00:02:01 yeah zipping sounds suspicious, probably some questionably legal activity 00:03:14 -!- alise has quit (Remote host closed the connection). 00:03:30 -!- alise has joined. 00:03:41 Suddenly 00:03:43 Astronomers 00:04:17 Everywhere 00:04:19 Note that it's not my dad claiming it, although he did decide he'd rather be safe than sorry 00:04:32 i think we've pretty much covered the "your dad is crazy" front 00:04:36 with glue 00:04:47 It's his AV, not him 00:04:50 SgeoN1: no but he _would_ choose that kind of AV, wouldn't he 00:05:01 McAfee 00:05:06 I think 00:05:28 It was a few minutes ago, why would I remember? 00:07:30 this would be a good time for a pun on really wanting to buy coffee, if intel hadn't run that into the ground already 00:07:37 McAfee AV is probably the worst consumer product ever released. 00:07:41 oerjan: wait, what? 00:07:51 -!- Guest76379 has quit (Read error: Operation timed out). 00:08:09 well reddit, commenting on intel buying mcafee. 00:08:13 Big Rigs... 00:08:35 oerjan: how coffee tho :| 00:08:40 Caffwe coffee 00:08:44 wait INTEL BOUGHT MCAFFE? 00:08:45 why? 00:08:49 McAFFE 00:09:10 My coffee 00:09:32 Hmm, I could take the place of the gone JokeExplainer 00:09:45 -!- yorick__ has joined. 00:10:07 alas, poor JokeExplainer. I knew him, barely. 00:11:43 -!- wareya has quit (Read error: Connection reset by peer). 00:12:41 -!- Sgeo has joined. 00:12:41 http://www.reddit.com/r/funny/comments/d3d8r/just_a_normal_day_for_an_intel_ceo 00:12:43 -!- wareya has joined. 00:12:52 it's in there somewhere, among the rest of the jokes 00:13:09 alise, will you watch me fail at Crawl now? 00:13:38 poor solaris 00:13:39 Sgeo: as soon as i finish messing this up sure 00:13:54 [[Intel CEO: "We need antivirus, can someone buy me McAfee?" Few hours later: "Done." "Great, which version?" "Version ... ?"]] 00:13:54 ha 00:14:01 old 00:14:06 heard it at least twice already 00:14:12 yes, well, I had no idea they'd even bought McAffe until three seconds ago. 00:14:20 it wasn't knew when it was on reddit either 00:14:22 *new 00:14:50 Antivirus programs are some of the worst consumer products ever released. 00:14:54 the McAffe misspelling should work so much better in german 00:15:04 how does gargoyle display better than my OS ... 00:15:08 McAfee is just worse than most others. 00:15:16 pikhq: nod32 is excellent on windows btw 00:15:32 you have to pirate it, but it takes up ~0 memory, is written entirely in assembly, and is one of the most effective for detecting shit there is 00:15:40 (1st or 2nd place) 00:16:05 pikhq: amen to that 00:16:38 "have to"? 00:17:02 Sgeo: well, or pay for it, I GUESS. 00:17:06 but who'd do that 00:17:36 -!- cheater99 has quit (Ping timeout: 265 seconds). 00:18:26 Ok, I'm back on Crawl 00:18:34 -!- relet has quit (Quit: Leaving.). 00:18:40 Sgeo: telnet address? 00:18:58 Not telnet, ssh 00:19:09 crawl.akrasiac.org user joshua password joshua 00:19:17 how do you use an arbitrary font in LaTeX? 00:19:24 bash: ssh: command not found 00:19:26 coppro: arbitrary what font? 00:19:31 coppro: TeX has its own font system 00:19:33 text font 00:19:36 I mean format. 00:19:37 * Sgeo mindboggles at alise 00:19:40 if you want OpenType fonts, you have to use XeLaTeX 00:19:41 which sucks 00:19:44 Sgeo: haven't installed it yet :) 00:19:57 coppro: so "google for a TeX version" basically. 00:20:01 there usually is one 00:20:04 + a supporting latex package 00:20:09 dammit, I was afraid of that 00:20:19 coppro: you /could/ use XeLaTeX... but it kinda sucks 00:20:26 alise: Or LuaTeX. 00:20:30 pikhq: yeah, but ... 00:20:32 why would you want to 00:20:51 Currently? No good reason; it's very much a work-in-progress. 00:21:17 In the future? Microtypography + OpenType. Need I say more? 00:21:34 Sgeo: you are impresent 00:21:39 pikhq: yeah but Lua :P 00:21:40 alise, Sgeoster 00:21:49 Sgeo: ah 00:21:49 because I forgot that I registered Sgeo 00:22:05 And I'm likely to die right now due to stupidity 00:22:27 alise: They're using it well, at least. 00:22:33 sgeo gets no presents 00:22:37 Sgeo: stop dying 00:22:40 Though I strongly suspect a Scheme would be better-suited. 00:23:11 alise, you're watching now? 00:23:15 yes 00:23:24 it's just like nethack but boring! 00:23:29 alise, boring? How 00:23:33 It automates boring stuff! 00:23:36 it looks boring 00:23:43 No price-ID, no Elbereth-spam 00:23:46 i'll have to write NetFuck 00:23:56 no Elbereth spam? but that's the whole fun thing about Elbereth 00:23:59 Automated travel between levels 00:24:05 E-ElberethE-ElberethE-ElberethE-Elbereth 00:24:22 ha ha sgeo dead 00:24:27 serves him right for betraying nethack 00:24:39 which level is that 00:24:41 ah, 4 00:24:56 alise, switching to the name Sgeo 00:25:45 Sgeo: psht, it's not as good as A Dungeon 00:25:56 ? 00:26:16 Demigod is a species 00:26:18 alise, playing again 00:26:32 Sgeo: A Dungeon is my now-named roguelike! 00:26:38 for alise dungeon, obvs 00:27:05 alise, you watching? 00:27:33 yes 00:27:45 5 is rest until something interesting happens 00:27:51 aka 100. 00:27:56 i have to be up at 9 :) 00:28:08 alise, should I start singing OpenBSD songs at you? 00:28:16 no 00:28:33 corpses aren't food 00:28:37 Unless they're cut 00:28:50 And the cut corpses won't be eaten unless hungry 00:29:03 And I wasn't expecting that 00:29:10 Kind of obvious though, really 00:29:44 corpses should be food... 00:30:41 Probably they're yelling at me for walking around with >100% health 00:31:27 Well, that was pathetic 00:31:57 And again... 00:33:48 Corpses are color-coded 00:33:55 The green means it's poisonous 00:34:09 it sounds too easy 00:34:35 Supposedly, spoiled NetHack is easier 00:34:46 Also, it's meant to not need spoilers 00:34:46 So 00:35:00 There's a lot of things the interface does for you 00:35:10 Mechanics are obvious, etc. 00:35:26 erm, obvious is the wrong word 00:41:31 There's an escape hatch in the ceiling 00:42:59 you ded 00:44:04 Sgeo: Should I call my roguelike Squirm, or something else? :P 00:44:11 Crawl, Squirm, ... 00:44:23 Only if it has a Crawllike UI 00:44:32 Otherwise, name it something more hacky 00:44:56 Whee, autoexplore 00:45:09 Sgeo: it has an it-like UI :P 00:45:32 Are you watching me? 00:45:52 yes, you're dead 00:45:55 oh 00:45:56 now you're not 00:46:55 Stupid cursed weapon... 00:47:48 Well, that was a waste of a magic mapping 00:48:13 And a blink 00:48:30 blink? 00:48:40 Short range controlled (if by scroll) teleport 00:48:46 Good for emergencies 00:48:51 why is it called a blink? 00:48:56 I have no idea 00:49:03 Well, most teleports in Crawl are delayed.. 00:49:06 Blink isn't 00:50:14 -!- Mathnerd314 has quit (Ping timeout: 272 seconds). 00:52:45 -!- wareya has quit (Read error: Connection reset by peer). 00:53:48 -!- wareya has joined. 00:54:33 ??Jessica 00:54:33 jessica[1/1]: Early unique human wizard who is very easy to kill, but between haste, slow, pain, blink, and possibly a wand you might not get a chance. Either a pushover or worse than Sigmund, depending on the will of the RNG. 00:54:38 OH NO JESSICA 00:54:52 THE SCARIEST OF NAMES 00:55:01 And she's dead 00:58:51 Check out my non-diagonal non-orthogonal aim! 01:00:23 alise, see this? 01:00:29 It means I'm going to teleport soon 01:00:31 okay 01:00:39 feel free 01:00:41 Check out my non-diagonal non-orthogonal aim! 01:00:42 how 01:00:44 also, *non-cardinal 01:00:55 alise, you weren't watching? 01:01:07 i mean how does that even work 01:01:17 oh, i see 01:01:28 Also, if there was an enemy nearby, it would autoaim 01:03:11 alise, it will autoaim, watch 01:03:46 Well, this seems unhealthy 01:05:17 alise, I'm going to make autoexplore skip that area 01:05:36 zoooom 01:05:44 As soon as I figure out how 01:06:10 There we go 01:08:13 GENTLY, ARCHAEOLOGY 01:14:43 -!- SgeoN1 has quit (Quit: Bye). 01:17:49 Sgeo dies 01:17:57 alise, what? 01:18:00 Nono 01:18:27 Ok though, that worm may be trouble 01:18:41 Anything yellow in inventory is an escape item 01:18:51 I have four scrolls of teleportation 01:18:58 They're not immediate though, so... 01:19:20 Oops 01:20:24 Nono <-- double negative, so yes? 01:21:48 Check out my non-diagonal non-orthogonal aim! <-- ? 01:23:39 In NetHack, effects, including ranged attacks, can only go up, down, left, right, up-right, up-left, down-right, down-left 01:23:52 Erm, screenwise I mean 01:34:12 alise, I die... 01:40:15 knife moves 01:40:19 *knight 01:40:21 ugh 01:40:38 it'd be interesting trying a knightspace renderer 01:57:23 Oooh, an altar 01:57:29 Sadly, altars don't do all that much 01:57:38 Erm no, they do 01:57:46 Just, I've been told to not do anything with them yet 01:57:54 Apparently, it's how you convert to a religion 01:57:58 And other stuff maybe? 01:58:02 Not BUC testing though 01:58:19 altar[2/2]: Unlike other roguelikes, you won't be sacrificing much - altars are generally only used for religious conversion. You also can't use them to discover curse status, sorry. Silly Hacker. 02:00:52 knight moves <-- right 02:01:03 it'd be interesting trying a knightspace renderer <-- hm? 02:01:31 Sgeo, of course you BUC test with an altar 02:01:42 and you can sac a lot 02:01:56 Sgeo, assuming nethack that is 02:02:37 YOu know what they say about assumptions... 02:02:42 Sgeo, asking rodney ?? altar 02:02:48 gives way too many lines 02:02:51 to paste here 02:02:55 Vorpal is retarded :p 02:02:58 Sgeo, you said " In NetHack, effects, including ranged attacks, can only go up, down, left, right, up-right, up-left, down-right, down-left" above 02:03:01 thus I assumed nethack 02:05:30 Sgeo, so which rougelike? 02:05:48 Dungeon Crawl Stone Soup 02:06:08 Vorpal: *roguelike 02:06:39 Sgeo, ah 02:07:27 Grr 02:07:36 Was wandering around without full magic again 02:08:58 My ghost killed someone else 02:10:39 Vorpal: i mean it'd be interesting trying to render a scene where you step by knight moves or something like that 02:10:48 afk, food 02:11:30 GreaseMonkey, render, as in 3D render? 02:11:56 knight geometry 02:21:50 -!- sshc_ has joined. 02:24:23 back 02:24:29 Vorpal: perhaps 02:25:22 I just want to slap coppro in here so that alise has a chance to either slap coppro or do some other alise-y thing that probably amounts to the opposite of a slap 02:25:22 -!- sshc has quit (Ping timeout: 276 seconds). 02:25:40 what 02:26:05 Sgeo: it's 2:21, i have to be up at 9, unfortunately, i am practically paralysed due to being creeped out 02:26:06 hooray 02:26:27 creeped out? 02:26:33 Hopefully not by what I just said :/ 02:26:34 reddit thread :P 02:26:38 http://www.reddit.com/r/AskReddit/comments/d4gu8/askreddit_what_are_some_unexplainable_things_you/ 02:26:57 Also, I can always sing you an OpenBSD song 02:27:02 I actually like the OpenBSD songs 02:27:54 i literally cannot move 02:27:57 so don't bother 02:28:18 Back when I was twenty/They said I wouldn't last/All that I believed in/Were the teachings of the past 02:28:38 All I ever wanted/Was to keep the world secure/And all the criticizing/Was something I'd endure 02:29:05 * Sgeo skips ahead a bit 02:29:21 Some say that I'm a hero/But I'm just being me/With my filter I can hide/My true identity 02:29:24 shut up 02:30:20 Welcome to the future/One very rich man/runs the Earth with/one multinational/owns your stuff/and owns your birth 02:31:11 Way back in my time/Open source kept/everyone choosing/People knew the insides/Of devices they were using 02:31:19 you're irritating 02:31:34 Go to sleep and I'll stop 02:40:22 yeah, not that easy when you're paralysed 02:40:38 Should I summon a newt to bite you? 02:41:28 shut up 02:41:30 :| 02:41:43 alise: my scam worked! 02:41:53 * alise pals coppro 02:41:56 (opposite of slap, Sgeo) 02:44:12 -!- oerjan has quit (Quit: Good night). 02:53:57 -!- sshc_ has quit (Ping timeout: 252 seconds). 02:57:20 Night all 02:59:06 night 03:06:17 -!- Sgeo has quit (Ping timeout: 265 seconds). 03:16:41 -!- cal153 has quit (Ping timeout: 272 seconds). 03:17:18 -!- sshc has joined. 03:38:28 -!- cal153 has joined. 03:52:49 -!- alise has quit (Quit: Leaving). 04:18:40 -!- Jubi has joined. 04:19:51 -!- Jubi has quit (Client Quit). 04:53:01 -!- sshc_ has joined. 04:54:25 -!- sshc has quit (Ping timeout: 276 seconds). 05:12:44 -!- sftp has quit (Remote host closed the connection). 05:46:03 -!- Mathnerd314 has joined. 06:08:21 -!- augur has joined. 06:10:55 -!- augur has quit (Remote host closed the connection). 06:11:01 -!- augur has joined. 06:18:54 -!- augur has quit (Remote host closed the connection). 06:18:59 -!- augur has joined. 06:31:36 -!- yorick__ has quit (*.net *.split). 06:31:37 -!- pikhq has quit (*.net *.split). 06:31:37 -!- coppro has quit (*.net *.split). 06:31:37 -!- fungot has quit (*.net *.split). 06:31:37 -!- mycroftiv has quit (*.net *.split). 06:32:50 -!- yorick__ has joined. 06:32:50 -!- pikhq has joined. 06:32:50 -!- coppro has joined. 06:32:50 -!- fungot has joined. 06:32:50 -!- mycroftiv has joined. 07:05:15 -!- Wamanuz has joined. 07:08:34 -!- FireFly has joined. 07:25:05 -!- pikhq has quit (Ping timeout: 252 seconds). 07:45:36 -!- augur has quit (Remote host closed the connection). 07:45:42 -!- augur has joined. 07:45:42 -!- augur has quit (Remote host closed the connection). 07:45:48 -!- augur has joined. 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 08:27:40 -!- sebbu has quit (Read error: Connection reset by peer). 08:28:04 -!- sebbu has joined. 09:12:39 -!- Mathnerd314 has quit (Ping timeout: 265 seconds). 09:13:35 -!- Mathnerd314 has joined. 09:17:42 -!- tombom has joined. 10:15:36 -!- wareya_ has joined. 10:18:21 -!- wareya has quit (Ping timeout: 252 seconds). 10:27:48 -!- FireFly has quit (Quit: swatted to death). 10:39:34 -!- augur has quit (*.net *.split). 10:39:35 -!- GreaseMonkey has quit (*.net *.split). 10:39:35 -!- comex has quit (*.net *.split). 10:39:35 -!- HackEgo has quit (*.net *.split). 10:41:00 -!- augur has joined. 10:41:00 -!- GreaseMonkey has joined. 10:41:00 -!- comex has joined. 10:41:00 -!- HackEgo has joined. 10:43:05 -!- ais523 has joined. 10:53:05 -!- comex has quit (Ping timeout: 276 seconds). 10:53:48 -!- comex_ has joined. 10:59:03 -!- tombom has quit (Ping timeout: 252 seconds). 11:17:56 both the C# camp and the Java camp are desperately trying to use this lawsuit to "prove" that their language is superior. 11:18:02 -!- oerjan has joined. 11:18:14 hmm, I wonder if we can convert everyone to esolangs like that? 11:35:57 A peculiarity: if you /part from our official state religion, any less than 12 year old children automatically follow, which I guess makes sense; but apparently you can also forcibly dis-join such children (without even asking them) even if you're still going to stay as a member yourself. 11:36:35 I wonder if any homes use this as a sanction; "you'd better behave or I'll kick you out of the church". 11:48:27 -!- GreaseMonkey has quit (Quit: New quit message. Entering 2006 in style.). 11:52:12 -!- MizardX has quit (Read error: Connection reset by peer). 11:54:12 -!- MizardX has joined. 12:00:23 ais523, what lawsuit? 12:00:29 Oracle vs. Google 12:00:42 ais523, oh? what is it they are suing about? 12:01:03 alleged patent infringement in Android 12:01:11 * ais523 vaguely wonders how Vorpal managed to miss this 12:01:17 it's been all over the tech news for about a week now 12:01:26 huh 12:02:00 ais523, btw did I mention that the ceiling in an auditorium at the university I attend fell down a few days ago? I had lectures in that very room this spring. 12:02:08 no, you didn't 12:02:17 that's sort-of embarassing for the ceiling manufacturer, actually 12:02:18 ais523, maybe you weren't in here yesterday then 12:02:23 I wasn't 12:02:33 people don't really want much from ceilings, but not falling is a major part of it 12:02:39 ais523, they are apparently no longer in business according to the local paper 12:02:51 was renovated in 1995 12:03:00 (when the ceiling was replaced) 12:03:14 There's a company called ZenRobotics, a sort of a spin-off of people from this university department and others; it was called Zendroid at first, but Google was all "you can't call it that, we have this Android thing". 12:03:41 See e.g. the latest newspost at http://www.zenrobotics.com/?page=news -- they have a rather unserious way of writing "press releases", if you can call 'em that. 12:03:55 ""We're not saying that Mika was key for Google's success story, and we're not alleging that Google's withdrawal from China is in any way related to this coup de grace from our side. Then again, there are many things we are not saying or alleging!", says ZenRobotics's CEO Jaakko Särelä. 12:04:02 "Well, actually I was the Lead Developer for the Latitude China team, but I had nothing to do...", starts Mika when Head of Special Ops Mr. Peltomaa drops him handily with a swift blow to the neck. "Mika has no more comments on the matter as he's apparently unconscious", sums up Mr. Peltomaa. The press is requested to leave the ZenRobotics office, amid rumors of harsh censorship." 12:04:16 (They hired a guy away from Google, which is what the newspost is about.) 12:04:40 fizzie, hah 12:10:31 -!- nooga has joined. 12:19:03 -!- oerjan has quit (Quit: leaving). 12:48:42 -!- Flonk has joined. 12:49:07 Hello. 12:49:21 hi 12:58:27 -!- tombom has joined. 12:59:01 hm is there any esolang that specifies an upper length for valid programs? And is of trivial computational class (like HQ9+ or such) 12:59:40 ais523, maybe you might know the answer to that? 13:00:15 I realised that you can implement some such esolangs in file(1) but it has to be even more limited than HQ9+ for that 13:00:17 I don't know of one 13:00:21 hmm, original Malbolge, maybe 13:00:30 ais523, flow control, too complicated 13:00:34 but it's more a bounded-storage machine 13:01:06 there must be a finite, and small, number of valid programs 13:02:01 hmm, so you can compile into a lookup table? 13:02:17 yes or submatches 13:02:33 something like h: output "Hello, ", w: output "world", u: output "user", would work with the additional clause that every instruction must only be used once in a single program 13:03:06 wouldn't need a full lookup table I think, some tricky submatches in file.magic would probably reduce the size of implementing it a bit 13:03:44 ais523, but yes, more or less you need to encode it in lookup table for the code, you can use values from the file though in the output. 13:03:59 Loops would be impossible unless you can look up all variants in a lookup table 13:05:39 You can do some pretty complicated things, according to man 5 magic. 13:06:27 Especially the test-type "indirect": "Starting at the given offset, consult the magic database again." And you can read the "given offset" from the file. 13:06:44 hm 13:06:50 maybe you could do loops then 13:07:15 fizzie, I wonder what computational class it is in then 13:07:58 There's also some sort of offset that's updated by regex tests that might let you do conditionals. I haven't written any "code" for file, and the man page isn't horribly clear. 13:08:31 More conditionals than just the "if true, output a message; if not, then don't" sort. 13:08:39 there is no memory though, so definitely sub-tc 13:12:25 Oh, right, there's that whole subtest hierarchy, so you can definitely do conditionals that do "indirect"-type test on different offsets depending on a main test. It's too bad there's only that one input file, and that one single offset pointer to it. 13:14:15 fizzie, hm "Offsets do not need to be constant, but can also be read from the file being examined." 13:15:02 seems you can get any constants from the file in memory, but arithmetics seems extremely limited 13:16:20 You can do the usual sort of arithmetics (+-*/%&|^) with one operand from the file, another a constant in the "source". 13:16:57 fizzie, yes but I still don't think you can really compute anything 13:16:57 But about the only sort of state you have is the offset to the file being examined. 13:17:12 exactly 13:17:30 fizzie, FSA perhaps? 13:18:47 I think it depends whether you want the magic(5) contents to be a fixed thing to execute a "program" in the file it's inspecting, or whether you're willing to allow a combination of specifically crafted file *and* a magic(5) "source" to be a "file(1) program" to compute something. 13:19:53 hm 13:20:13 fizzie, the latter sounds really hard to classify 13:20:31 Even in the latter case it sounds pretty tricky to do anything nontrivial. 13:20:39 well yes 13:20:42 -!- sftp has joined. 13:22:43 But, for example, if you make a file that has the bytes 0x00 0x01 0x02 .. 0xff, in order, then your magic(5) file can use the "current offset" value directly in the tests, since the value at the offset equals the offset. (Of course with the offset limited depending on how large a file you want.) 13:23:23 heh 13:24:48 fizzie, I can see how you would do HQ9+ except for Q and +... H9 hm... extremely silly language 13:25:30 the issue with Q is that while you could do it, it would mess up the offset 13:25:41 so you probably no longer have any clue where to continue the program 13:30:37 I'm not sure about that. The relative-offset (&0) tests are relative to "the last up-level field", so if in a ">>&0 string Q" you detect the command, in a ">>>"-level test you could then print out the complete file, if it's possible to do so with one test. Still, I haven't completely thought this through. And even HQ9 is still pretty boring. 13:31:41 -!- Mathnerd314 has quit (Quit: ChatZilla 0.9.86-rdmsoft [XULRunner 1.9.2.8/20100722155716]). 13:32:28 fizzie, hm 13:32:42 fizzie, you will need a indirect match to continue though, no? 13:34:28 No, I think you just need to do something like ">&1 indirect" entry, where that's on a level where the Q execution hasn't messed up the correct value of "&n"; I think "&1 indirect" should mean "starting from the next byte, consult the database again". 13:34:56 I'm not sure how to make it terminate; I don't know what values the tests "return" if you refer to outside the file. 13:36:09 It might even be that the "consult the magic database again" means that on the second run, constant-offset "0" would refer to what was the parameter of indirect. It's not very well documented in this manpage. 13:40:35 It seems to, which would make "Q" very hard to do in the simple architecture where you'd continue with ">1 indirect" after each matched HQ9 command. Because I don't know how you could get to the real beginning of the file to print it out. 13:41:51 -!- pikhq has joined. 13:41:55 Right, if you just do a 0 regex .* %s, it will only print the half of the file starting from the Q. 13:43:00 http://sprunge.us/ZbaY 13:43:30 Maybe if you just respec the Q command to be a "half-quine"; at least it's novel and new and other synonyms. 13:43:58 (There's some newline issues in that example, but those are minor.) 13:52:37 hm 13:52:45 fizzie, nice 13:53:34 fizzie, it prints out ? 13:53:59 Yes, I'm not sure how to stick a literal newline in there, if it's possible at all. 13:54:30 fizzie, you could make the interpreter command a bit more complex, like appending sed 's//\n/g' or such 13:55:07 or hm 13:57:57 You can replace the "0 regex .* %s" with the potentially more appealing "0 string >\0 %s"; I think both have the problem that they don't match across newlines, but you can just strip those from the hq9 source. Still, the "half-quine" problem is a worse one. 13:59:35 hm 14:01:49 If you use "indirect" to restart the processing, I'm not sure how to get around that. If you move the "starting offset" forward to use it as an IP, it doesn't seem possible to get back to earlier parts of the file; but if you do "indirect" with a start offset of 0, it is as if you'd start the whole thing from the beginning, since there's no state you could use to distinguish from where you were. 14:04:00 fizzie, 99 bottles seems to break it hm 14:04:08 Oh, I didn't add 9 in at all. 14:04:11 a few lines of it work, but more and it prints it incorrectly 14:04:14 fizzie, yes I added it 14:04:31 Maybe it has some length limits in there. 14:04:39 hq9.magic, 3: Warning: description `99 bottles of beer on the wall, 99 bottles of beer.Tak' truncated 14:04:49 damn 14:05:27 hq9.magic, 4: Warning: offset `l, 42 bottles of beer.Take one down and pass it [...] Go to the store and buy some more, 99 bottles of beer on the wall.' invalid 14:05:46 Does it happen if you do it with multiple messages? Something like "0 string 9" followed by ">0 default x first line of song", ">0 default x second line of song", ...? 14:06:03 ("offset default x" should be a test that always matches.) 14:06:37 ah could try that 14:07:52 It might still truncate the full message, though. 14:08:17 fizzie, it seems to not print the default one 14:08:32 wait, maybe I'm doing it wrong 14:08:42 ah yes needs the x test 14:09:04 Yes, "default" is just the data type. 14:10:53 hm 14:11:05 Take one down and pass it around, 98 bottles of beer on the is the max length hm 14:11:15 62? 14:11:16 huh 14:17:50 Apparently if you use an indirect offset with the indirect type -- something like ">(0.b) indirect" -- it ignores the current starting offset from the previous "indirect", and instead uses the offset read from the file directly. So if you have a 0 there, you'll get back to the real beginning of the file. But it still doesn't help, since there's no state, and when you're back at the beginning of the file, it's exactly identical as when you first ran it; so you c 14:17:50 an't go back to where you were. 14:18:03 Also, there's a bus I need to catch; away for now, back late in the evening. -> 14:24:33 hm 14:26:11 Assuming you can do a relative indirect offset (I think you can), and are allowed to do a simple transformation, which unfortunately will take N*S*A of size (where N is the length of the input, S is the number of states, and A the size of the input alphabet), I think you can do a FSM. :p 14:27:35 (Basically, you'd use the current offset to denote both the position in input as well as the state -- giving N*S places -- and in each place put the transition table of that state, which of course has A entries.) 14:29:26 Of course anything capable of comfortably building the special input file would probably be more than enough to directly do the FSM itself. But at least the magic db would be very simple, and generic. 14:57:30 -!- ais523 has quit (Quit: Page closed). 15:21:13 -!- FireyFly has joined. 15:30:06 fizzie, current hq9 file: http://sprunge.us/BNUe 15:30:19 that is the magic 15:31:03 hm there seems to be an issue with adding a space at the start of lines 15:31:38 ah yes 15:31:42 need to move about 15:38:28 fizzie, here is a new version http://sprunge.us/gXJf 15:38:34 including the wrapper script 15:51:22 fizzie, hm it might be able to do Q 15:51:29 fizzie, by heavy preprocessing 15:52:44 basically insert a the length to end of file after ever Q pointing to the end of the file (after taking these length fields into account of course), then append the original source, That way you could get the original source by using that offset 15:53:37 fizzie, that ammount of pre-processing makes it "not really file(1) any longer" though 15:54:21 hm you could mark it with a # (you need to know when to stop reading the file anyway, so you need a marker somewhere) and then use search and skip the length value 15:59:16 -!- Phantom_Hoover has joined. 16:00:40 -!- FireyFly has changed nick to FireFly. 16:01:45 -!- Sgeo has joined. 16:06:49 fizzie, hm 16:06:57 this seems to work assuming no newlines in input file 16:07:06 could of course escape those 16:10:21 15:28:50 Near as I can tell, the *simplest* way to go to college in the US without being in debt forever is to get married. ← a solution appears immediately. 16:11:32 Phantom_Hoover, married? how so? 16:11:45 I dunno, pikhq said it. 16:11:52 Or how to get married? 16:12:11 as in "why would getting married help avoid debt?" 16:12:11 That's easy enough if you can find someone of the opposite sex who also wants into college. 16:13:10 Well, in the logs he said it was because they calculated things based on your own income and assets rather than with those of your parents added. 16:13:48 fizzie, this one works assuming no | in the input source file: http://sprunge.us/NKiR 16:14:05 Phantom_Hoover, ah 16:14:30 I think in Scotland university is actually free for Scottish students. 16:15:19 Or with little to no fees. 16:15:25 in Sweden it is free for students from EU (or was it from Schegen maybe?) 16:15:36 Perhaps. 16:15:38 Schegen? 16:16:04 like EU + Norway, Switzerland, Listenstein and one or two other countries 16:16:36 it used to be free for students outside EU too, but this is the last year of that. 16:18:18 Phantom_Hoover, ah it is "Schengen" 16:18:21 missed a letter there 16:18:31 http://en.wikipedia.org/wiki/Schengen_Area 16:38:48 -!- MigoMipo has joined. 16:56:39 -!- leBMD has joined. 16:57:07 greetings, #esoteric. 16:57:20 Greetings, leBMD. 17:01:35 -!- derdon has joined. 17:02:15 So, how about them yankees? 17:03:05 What about them? 17:03:12 Ooh, graue replied to my email. 17:03:13 I have no idea. 17:05:20 Phantom_Hoover, nice, was he going to enable that thing? 17:05:29 or is* 17:05:32 I guess 17:05:35 He said he'd look into it. 17:05:55 And that if he didn't respond for a month, to email him again. 17:06:03 haha 17:07:10 So, here's a random poll: what languages are you guys currently doing? 17:07:38 when I say "currently doing" I mean "maybe messing around every once in a while" 17:07:54 Eso or normal? 17:07:59 Eso. 17:08:13 Well, I did some stuff in Lazy K a while ago. 17:08:50 -!- zeotrope has joined. 17:09:35 Cool. 17:09:48 leBMD, do you count "implementing interpreters for" as well as "programming in"? 17:09:56 sure. 17:09:57 RIP Sgeo 17:10:37 I've been doing Befunge, along with contemplating ways to make a Noobinary interpretor. 17:10:41 befunge98, brainfuck, INTERCAL and a few more. Special for today is implementing HQ9+ in file(1) 17:11:29 Vorpal, do tell. 17:11:29 while file(1) is in no way esoteric, this usage of it is 17:11:55 Phantom_Hoover, do tell what? 17:12:19 The nature of your HQ9+ interpreter in file. 17:12:21 wow. 17:12:25 -!- cheater99 has joined. 17:12:28 Phantom_Hoover, well it is a joint effort with fizzie really 17:12:37 fizzie, this one works assuming no | in the input source file: http://sprunge.us/NKiR 17:12:42 yeah needs a wrapper script 17:12:45 otherwise Q won't work 17:13:04 since the only state you have is your current offset, and you need to read the whole file somehow for Q 17:13:13 of course + is a dummy-implementation due to this 17:14:57 leBMD, what was that about noobinary? 17:15:11 I'm thinking about making an interpreter for it. 17:15:30 It wouldn't be too hard. It would just need a 1D vector for the stack, and a couple other things. 17:15:36 Indeed. 17:15:50 Oh, god, DMM is going on a pun craze... 17:15:51 quick fact: I'm the one who made noobinary. 17:16:06 ...which is probably why it's not very creative. XD 17:16:55 The comp. class looks to be PDA, at best. 17:18:48 leBMD, noobinary? 17:19:04 Vorpal, you know, the wiki has got a search function. 17:19:08 Yes, it's in the language list. It's made to resemble binary, but be easier to read. 17:19:15 A crappy search function, yes, but it's enough. 17:21:44 "Noobinary is an esoteric language designed by user Batmanifestdestiny to resemble Binary, but be easier to learn. " <-- Binary does not seem to be another esolang hm. 17:22:00 * Phantom_Hoover assumed he meant machine code 17:22:01 so what is it referring to? The number system? 17:22:24 I mean, we speak of binaries. 17:23:02 such as jpeg images yes 17:23:03 :P 17:23:07 Binary, other than the numbering system, is the very basic core of programming. 17:23:18 Not really true... 17:23:21 Phantom_Hoover, I tend to say "programs" "executables" "object files" or such 17:23:56 Well, considering it's the series of electrons going through the CPU, I'd say it's pretty basic. ;) 17:24:05 leBMD, binary is just the way machine code is represented. 17:24:27 The lowest level of abstraction is effectively assembly. 17:24:29 Phantom_Hoover, maybe I use binaries in the sense "unknown/arbitrary digital data on a binary computer" but that is it. Oh and the erlang data type binary, which is exactly that, arbitrary byte-stream that you can interpret however you want 17:24:46 well byte stream is the most common source/destination I guess 17:25:03 it is not a stream in the sense of an open FILE* in C 17:25:05 I would define it as a file not intended to be readable with a flat text encoding. 17:26:05 Phantom_Hoover, hm so how do you classify that uudecode *.com that consisted of only printable chars? 17:26:12 I think ais linked to it a few years ago 17:26:32 iirc he (or maybe someone else, don't remember) wrote it for sending over usenet, to people lacking uudecode 17:26:37 Well, probably. 17:26:51 Phantom_Hoover, by your definition it is not a binary? 17:26:57 Vorpal: Given that the file offset is the only thing that could even concievably be called "accumulator" there, and + increments it (as do all other commands, of course), it's a reasonable implementation. 17:27:08 fizzie, hah 17:27:15 fizzie, still the issue of | in the input file 17:27:27 fizzie, would need a bit more complex escaping to handle it properly 17:28:15 Oh gosh, my cat is meowing in her sleep. 17:28:21 The "include source with marker" is also a bit cheaty, but, well, understandable also. 17:28:23 leBMD, has this conversation been illuminating? 17:28:37 Sure. 17:28:53 fizzie, I think we more or less exhausted all possible alternatives to it 17:31:51 If you set an upper limit to program length, you can so something inelegant like including the H, Qa9 17:31:56 fizzie, any progress on jitfunge btw? 17:32:05 and 9 handlers separately for each offset. 17:32:21 Nothing to report on that front. :/ 17:32:34 Surely you can do JITfunge with an existing compiler with a hook stuck onto the p instruction? 17:32:47 um 17:32:50 Phantom_Hoover, what? 17:33:19 Compile it, stick the result into a buffer, call, when p is executed, recompile. 17:33:43 There aren't that many existing befunge compilers either. 17:34:00 Phantom_Hoover, how would you handle ~~x then? 17:35:01 What does x do 17:35:03 yes an extreme example, but x and j are both problematic even without parameters taken directly from user input 17:35:13 Phantom_Hoover, pop dy and dx and set the delta to that 17:35:13 And indeed j. 17:35:30 Ah. 17:35:36 and j pops a parameter n then jumps forward n*current_delta, n might be negative 17:35:41 and indeed 0 17:36:10 Flick to interpretation until the delta has settled again? 17:36:20 Or just do B93? 17:36:24 Phantom_Hoover, befunge93 you could feasibly compile all 4 cardinal directions for, but 98? no way 17:36:44 since it has arbitrary delta thanks to x 17:36:55 Phantom_Hoover, and jitfunge is 98... 17:37:05 Bah 17:37:12 Phantom_Hoover, so 93-only is not really relevant for this discussion 17:38:15 Phantom_Hoover, besides you might think implementing + as straight forward, but then you forgot that popping on empty stack pops 0. So you need to check for that 17:38:43 Recompiling absolutely everything on each change doesn't sound so incredibly clever. And I'm not claiming there to be any sort of theoretical reasons why JITting befunge would be somehow incredibly difficult, just that it is not entirely trivial either. 17:39:06 fizzie, btw, what is that strangeness you do with mmap? IIRC recent linux introduced a guard page below the stack and above the heap. Recent = 2.6.35 17:39:16 not sure if this might mess up for you 17:39:22 but it made lvm spit warnings 17:39:39 need to reboot to 2.6.25.3 soon, it has a fix for lvm doing that 17:39:48 (At least if you want to (within limits) minimize the amount of needless recompilation.) 17:40:27 Dunno, I mmap in the middle of nowhere and put my own guard pages around the b98 stack, I don't see why that should be a problem. 17:40:41 fizzie, I didn't say JITing was increadibly difficult. What I suggested was that Phantom_Hoover's approach to the problem was infeasible. 17:40:56 Indeed 17:41:00 that said, JITing is _quite_ difficult at least. Even for non-befunge 17:41:14 fizzie, right 17:43:10 The mmap-segfault-autocheck stack is broken anyway with the llvm codegen at the moment, since the handler needs to understand all the different operations that could try to access the stack (to fake the stack pointer as if a zero was returned without it moving), and you never quite know what the llvm code will turn into. It might need some more llvm-digging. 17:44:48 I optimize pop+push -> peek, which also complicates it a bit, since a "peek" operation on empty stack should actually do push 0. 17:45:49 Not that many programs probably care whether there's a "real" zero on the bottom of the stack, on top of the imaginary ones. But it certainly has an observable effect. 17:46:46 fizzie, mycology cares (it checks with y) 17:48:03 Mycology is a bit special that way. I mostly meant the huge amount of B98 applications that are used all around the globe and ha wouldn't it be nice if there actually were any? 17:48:16 Besides fungot, anyway. 17:48:17 fizzie: you can not implement scheme using a java irc library :d) there just would be kewl.). it has awk macros, a wonderful typo, imho) 17:48:54 fungot, correct. 17:48:54 Phantom_Hoover: it's very easy to get the interface of the sicp 17:49:04 Of course. 17:54:12 fungot is so cool 17:54:13 Phantom_Hoover: got a microphone? i might need a bit of ratpoison advocacy slip out there. similar logic applies to physical theories: that's why you agree to these patches going into the kitchen and starts making a lot of 17:54:55 Cool but sometimes a bit incoherent. 17:55:10 He's coherent locally 17:55:33 This might be an old thing, but from #anotherchannel: http://bit.ly/9DBmOT -- (a maps.google.com link) 17:58:07 Zuh? 17:58:19 -!- relet has joined. 17:59:02 It looks as if it is achieving warp speed. 17:59:37 Oh, wait, I get it. I think. 17:59:53 I assume the different colours are captured sequentially? 18:00:21 And the brightness slightly later, by the looks of it. 18:00:49 I don't know about brightness, but it does look like they take R, G and B pictures at different moments. 18:01:26 One would think that "brightness" comes from the sum of it, though; it's not like they could physically speaking capture the chrominance and luminance information separately in any sensible way. 18:03:00 Well, there's the grey shadow after everything else. 18:04:21 Oh, right, that. Yes, I think they could be taking one more "monochrome" picture with no filters to get better luminance resolution, or some such thing. 18:04:57 I had managed to miss the "outline" part completely somehow. 18:33:39 "Parts of its functionality are almost copy-pasted from CCBI, and I think AnMaster trusts CCBI a bit more than he should. In some cases, this means that CCBI bugs remain in it for some time. But that does help it in getting a full pass from Mycology." ← is this true? 18:38:19 -!- Flonk_ has joined. 18:39:46 -!- Flonk has quit (Ping timeout: 255 seconds). 18:40:01 -!- Flonk_ has changed nick to Flonk. 18:43:02 Phantom_Hoover, I believe I did more or less copy parts of 3DSP at that time since I had not yet learnt matrix math. Apart from that: no, though in cases were the standard were unclear I used it as a reference implementation. 18:43:29 Ah 18:43:36 Phantom_Hoover, so I would call it quite an exaggeration. 18:44:24 Well you would say that! It's just like the evolutionists! 18:44:28 Phantom_Hoover, however, I did credit him where credit was due, and one of the point of open source is sharing :P 18:45:02 points* 18:46:08 And "sharing" is just another word for "stealing"! (They even begin with the same letter.) 18:46:41 omigodrobotwarsisbeingrerun 18:47:42 Is this the I-consider-it-just-a-fake remote-controlled thing? 18:47:47 Indeed. 18:48:02 But the DESTRUCTION! 18:48:05 It's not real "robot wars" unless autonomous, amoral killing machines are involved. 18:48:19 Well, let's do that! 18:49:41 del morality 18:49:47 (Python) 18:50:05 why can't they just make them autonomous? 18:50:18 it would make it way more interesting 18:50:39 It would, but that would require decent programmers as well as mechanics. 18:50:52 Or it could make it way more boring, given how dull e.g. robot soccer is. 18:50:54 who is fizzie 18:51:04 cheater99, the writer of fungot? 18:51:04 Phantom_Hoover: can you credit me... no problem for a fnord 18:51:21 Ow, that's quite a thing to be known for. 18:51:22 dunno 18:51:23 My god, we need to make him into a fortunealike 18:51:30 i'm just looking at the fizzie nick 18:51:40 and i can see it does not belong in here 18:51:43 it's an intruder 18:52:09 i think i should go to the city or sumptin 18:52:12 or at least to the shop 18:52:43 Oh noes, I am INTRUDING. Perhaps even EXTRUDING. 18:53:01 * Phantom_Hoover sends our autonomous killing machine to get fizzie. 18:53:26 If they could self-repair, then it would become awesome^2. 18:58:11 It would, but that would require decent programmers as well as mechanics. <-- um yes? 18:58:34 I fail to see the problem with that 18:58:39 It raises the bar greatly. 18:58:51 Phantom_Hoover, how is that a bad thing? 18:59:09 Because you can have about one competition a year. 18:59:17 yes and? 19:00:08 There's less to watch. It's about entertainment, after all. 19:00:30 Phantom_Hoover, expand the area of coverage, then split it up in loads of sub-matches 19:00:36 they manage to do that with football after all 19:00:41 and that is ever 4th year iirco 19:00:42 iirc* 19:00:44 or was that OS? 19:01:10 Vorpal, these things are very complex; custom stuff, etc. 19:01:34 Designing strategies would be much harder than for football. 19:01:35 Phantom_Hoover, football is also very complex, I never managed to understand what "offside" is. 19:01:42 Also, and this is just a guess, but I doubt they'd be so free about installing the kill-and-maim hardware I assume those things have (I don't really follow any of the shows) if they weren't just remote-controlled toys. 19:01:49 Vorpal, true, but after that... 19:02:02 Phantom_Hoover, after that I gave up on football so no clue 19:02:12 The rules are static, the hardware is basically the same and there's not many possible strategies. 19:02:37 Phantom_Hoover, sounds like football? 19:02:45 hardware = the ball ? 19:03:08 For a destructive arena game, you have a huge range of opponents, you have unique hardware. 19:03:15 Phantom_Hoover, but you could of course make rules more flexible, if you make the bot autonomous you have to change the rules anyway 19:03:28 You have to be self-correcting, in case a wheel gets ripped off. 19:03:42 of course, or use a different traction system 19:03:43 You have to be able to adapt strategies for each opponent. 19:04:18 You have to take into account the state of the opponent, since they might themselves have a wheel ripped off. 19:04:19 Phantom_Hoover, so it is more like judo or perhaps boxing? 19:04:26 destructive and so on 19:04:34 No, they are still rather easier. 19:04:49 There are a huge range of variables here, all of which have to be coded. 19:05:03 I'd still watch it, even with imperfect autonomous killer machines, but I do think they get participants easier this way. 19:05:24 And you can't just stick a supercomputer into it; your computer needs to be inside the machine and needs to be shielded. 19:05:25 Phantom_Hoover, self adapting robots do exist 19:05:36 Vorpal, in experimental and fragile states. 19:05:36 Phantom_Hoover, so you don't have to code for absolutely everything 19:05:43 hm 19:06:10 You couldn't use plastics or anything, they'd be ripped to shreds by a nice, big saw. 19:06:38 Phantom_Hoover, also, modern computers are quite powerful. Even a mobile CPU 19:06:47 and you don't exactly need 8 hours of battery life there 19:07:16 Yes, but you're also running a large lump of metal with drive systems and armour. 19:07:57 true, hm... Why not allow the program to run from a remote system, but once the match started it is hands off? 19:08:01 Although you could have a human "mission control" who can adapt the overall strategies on the fly. 19:08:11 and yes that could be allowed 19:08:21 Vorpal, again, suppose a tyre or something is damaged? 19:08:38 Phantom_Hoover, guess why tanks do not use tyres! 19:08:51 Phantom_Hoover, why not use a similar traction system to that 19:09:10 Still vulnerable, and you miss the overall point. 19:09:19 Phantom_Hoover, it would however be way more cool 19:09:22 And again, the entertainment value: look at this RoboCup video at http://www.youtube.com/watch?v=ICgL1OWsn58 -- I guess it could be considered interesting to look at, but exciting mainstream TV it isn't. 19:09:37 -!- sftp has quit (Remote host closed the connection). 19:09:37 Phantom_Hoover, and correcting for damaged drive isn't terribly difficult 19:09:40 i.e. in a fight, you are likely to sustain damage that makes your old tactics ineffective. 19:10:09 Phantom_Hoover, so that is where mission control comes into it 19:10:17 -!- sftp has joined. 19:10:21 Indeed. 19:10:32 That could actually be workable, though still difficult. 19:11:47 Phantom_Hoover, anyway you basically want armour, lots of it traction a bit inwards from the sides, low GC, some way to flip over if you end up upside down (not hard) and some weapon. 19:11:49 right? 19:11:57 GC? 19:12:03 Center of gravity? 19:12:08 Garbage collection? 19:12:09 yes 19:12:11 the former 19:12:16 Anyway, yes, those are basically the requirements for the existing things. 19:12:20 often called GC in aircraft context 19:12:25 sorry for the jargon 19:13:00 did you know that the Concorde pumped fuel during flight to change the GC along the forward/aft axis 19:13:00 OK, now for the software. 19:13:14 primarily to avoid the drag from trimming the normal way 19:13:40 Wait, no. 19:13:50 (the normal way would be to angle the control surfaces from the "flat" position, but that induces drag) 19:14:03 Phantom_Hoover, and the software for that wouldn't be too complex 19:14:22 What about sensors etc? 19:14:41 Phantom_Hoover, well true, but sensors are light compared to the rest 19:15:03 True. 19:15:27 Phantom_Hoover, and power usage is also rather low compared to the movement 19:16:01 Can they be mounted so that they're fairly resilient? 19:16:04 * Phantom_Hoover → food 19:16:14 Phantom_Hoover, well the gyro and such definitely can 19:16:25 and camera probably could too 19:16:32 Phantom_Hoover: You can transform into food? 19:16:44 Well, I guess people are technically speaking food too. 19:16:53 was just about to say that 19:17:05 anyway 19:18:30 -!- oerjan has joined. 19:18:44 a small IR camera in each direction wouldn't be too hard. it would take a lot of luck for the opponent to hit a recessed square of hardened glass about 0.5x0.5 cm 19:18:54 Phantom_Hoover, remember that the opponent also is AI controlled 19:18:59 which actually lowers the bar 19:19:42 No, it just means that both opponents can't really fight very well, which makes for non-flashy battling. 19:20:02 fizzie, indeed, but it is a lot more interesting of course 19:20:09 Interesting for us, maybe. :p 19:21:08 Phantom_Hoover, you would need gyroscope or just some simpler way to detect being upside down, such as this: 19:21:09 ____ 19:21:10 __/ 19:21:10 ____ 19:21:11 see? 19:21:22 that is in the upside down position, would fall down 19:21:41 and if you built it the right way, being on edge is unlikely 19:21:47 so you don't even need a gyroscope 19:22:23 you do however need some odometer, but that is cheap and trivial to attach to the drive, and some position sensing of your weapon (depends on what sort of weapon) 19:24:16 Phantom_Hoover, hm... I would suggest using some titanium alloy for the armour... Quite light, strong, and not very brittle 19:24:46 Making sense of the real world even badly (well, for some values of "bad" that still let you do at least *something* sensible) is pretty nontrivial, though. I guess for a fixed-environment arena you could get something done with a drive-odometer (though wouldn't the derived position information start to drift there too?), but understanding what the opponent is doing is going to be pretty tricky. 19:25:16 -!- augur has quit (Ping timeout: 276 seconds). 19:25:26 probably with a metal skeleton below 19:25:53 fizzie, yes I was getting to that 19:26:20 fizzie, I suggest a laser 3D scanner if that is feasible 19:26:23 it's an intruder <-- that's quite something to say about the channel's top present admin... 19:26:27 fizzie, otherwise some IR camera 19:27:00 that takes care of where the opponent is 19:27:12 and perhaps also what it is doing 19:27:14 of course there _could_ have been a hostile takeover involved. 19:27:16 -!- Sgeo has quit (Ping timeout: 265 seconds). 19:27:27 oerjan: You're on the same level in the food chain, aren't you? 19:27:36 Oh, but you weren't here at that moment. 19:27:56 fizzie: i am? wasn't it you who made me op? 19:28:12 I don't remember doing that, but I could be wrong. 19:28:18 whaaat? 19:28:19 Vorpal, re armour etc., these details have been worked out. 19:28:21 -!- leBMD has quit (Quit: ChatZilla 0.9.86 [Firefox 3.0.13/2009073022]). 19:28:26 They don't change for us. 19:28:30 hm in that case maybe it's lament who is the top present admin 19:28:44 Phantom_Hoover, titanium alloys with supporting steel skeleton seems like the best bet 19:28:49 -ChanServ- 2 fizzie +votsriRfA [modified ? ago] 19:28:50 -ChanServ- 3 lament +votsriRfA [modified ? ago] 19:28:50 -ChanServ- 5 oerjan +votsriRfA [modified 26 weeks, 3 days, 21:05:35 ago] 19:29:07 We all have the same flags. (I don't know how the numbering is determined, though.) 19:29:15 it doesn't say who did it, though... 19:29:17 votsrirfa! 19:29:44 Vorpal, look it up. The RC models will do it better than you can make up in an afternoon. 19:29:45 !flags 19:29:47 err 19:29:49 right 19:29:58 chanserv not in here and freenode doesn't use that module anyway 19:29:59 hm so does that mean i could actually lift someone else up to my level as well? i thought i was some kind of under-op 19:30:01 so that won't work 19:30:24 Under-dog. But yes, I think that's all the flags you can have, except the "founder" flag that andreou has. 19:30:25 oerjan, yes you have +f 19:30:30 oerjan, /msg chanserv help flags 19:30:31 for details 19:30:41 ok then 19:30:55 * oerjan cackles evilly on principle 19:31:03 hah 19:31:07 Vorpal, so, some general strategy things. 19:31:27 Phantom_Hoover, titanium might be a bit expensive though 19:31:28 hm 19:31:30 -!- Sgeo has joined. 19:31:40 Hmm... perhaps robots could broadcast some information on their position and location. 19:31:52 Phantom_Hoover, just use LIDAR 19:31:54 To make things a little more interesting. 19:31:59 -!- yorick__ has quit (Quit: Reconnecting). 19:32:02 -!- yorick has joined. 19:32:26 -!- yorick has quit (Changing host). 19:32:26 -!- yorick has joined. 19:32:39 Vorpal, look, making the robot ascertain complex details like direction and identity from the poor sensors you have suggested is going to be really difficult. 19:33:08 Phantom_Hoover, Laser gyro should take care of it better than anything else 19:33:15 Phantom_Hoover, for position and orientation 19:33:32 For its own position and orientation, not that of the others. 19:33:43 Phantom_Hoover, of course you could use 4 radio beams, one in each corner of the arena, and then just do what GPS does 19:33:45 from those 19:34:00 a lot simpler than the full thing a GPS does 19:34:17 Yes, but then you have to know where your opponents are. 19:34:20 you don't have to wory about terrain, elevation, multi-path (well maybe that to some degree) 19:34:22 And where they are going. 19:34:32 Phantom_Hoover, that is where radar or lidar or similar comes into it 19:35:01 -!- yorick has quit (Client Quit). 19:35:04 -!- yorick has joined. 19:35:32 -!- yorick has changed nick to Guest86443. 19:35:46 -!- Guest86443 has quit (Client Quit). 19:35:55 Phantom_Hoover, or why not a sonar? 19:36:22 I am a robot. I want to know where the other robots are, what direction they are pointing and how fast they are going. 19:36:40 Phantom_Hoover, radar, lidar and sonar can all answer that 19:37:04 Phantom_Hoover, directly for position, speed and direction by integrating over time 19:37:14 If it's standing still? 19:37:54 And what about the flames shooting out of the ground, and holes that sometimes open there? (I looked at this first youtube-hit for robot wars -- http://www.youtube.com/watch?v=uA_z51dlPQU -- and the arena is a lot unfriendlier than your standard robot-soccer field.) 19:38:02 Phantom_Hoover, well, then you could try to do image processing to figure out the shape. You could load 3D models of the opponents in advance 19:38:18 Phantom_Hoover, or why not require them to have a blue spot on the front 19:38:19 or such 19:38:33 well then you need visual camera, no big issue 19:38:51 fizzie, I fail to see the point of those flames 19:39:13 Hah! Image processing objects out of grainy, noisy, messy, blurry visual images is *really* something that your average remote-controlled robot enthusiast is not going to be able to write. 19:39:31 DMM does image processing... 19:39:38 19:39:54 fizzie, well okay, so why not require it to be painted in radar absorbing paint at the front? 19:40:01 well not the whole front of course 19:40:04 just a small spot 19:40:04 Vorpal, please watch that video. 19:40:10 Because that's a lot more boring? 19:40:13 That's basically what we're aiming for. 19:41:07 Or what I was. 19:41:10 And again, personally I'd watch a lot less visually flashy autonomous-robot murder-o-death-o-matchery, but not everyone would. 19:41:11 I don't know about you. 19:41:30 Yeah, we can go without flames and pits. 19:41:47 Though some degree of environmental control could be interesting. 19:43:36 Phantom_Hoover, the sound quality is horrible 19:43:43 I have no idea what they said really 19:43:46 couldn't pick it out 19:43:49 It's a terrible video. 19:43:51 Ignore it. 19:43:54 Phantom_Hoover, and increadibly boring 19:44:08 Phantom_Hoover, ... you told me to watch it above 19:44:11 Vorpal, please watch that video. 19:44:19 Yeah, for context. 19:44:24 Anyway. 19:44:46 Phantom_Hoover, extremely boring 19:45:00 OK, so software design? 19:45:06 yes that is the interesting part 19:45:30 Yeah, they should be reviewing the code for both bots before the match! That's sure to hook in the viewers. 19:45:35 the people building the hardware would need to team up with people working on software, unless they are skilled at both 19:46:01 fizzie, reviewing hm? Why? 19:46:09 Because it's the interesting part, of course. 19:46:12 hah 19:46:44 fizzie, they could publish it as bonus material on their site, and then discuss the interesting parts on the show 19:47:11 oh and the sound level needs to go down. any sport where audience is louder than in Tennis is basically not interesting 19:47:16 Vorpal, this is going to be the nerdiest thing ever. 19:47:17 tennis* 19:47:23 not sure why I hit shift there 19:47:28 Phantom_Hoover: Nerdiest non-existent thing ever. 19:47:31 Phantom_Hoover, so we have a win-win situation! 19:47:42 fizzie, you have no AMBITION 19:47:58 Phantom_Hoover, and of course they could review the hardware as well as the software 19:48:04 both parts are interesting 19:48:05 No, but I have an amb-eval here somewhere. 19:48:19 Vorpal, to US. 19:48:40 But by all means, do make it happen. I'll even watch it, if it's watchable over the interwebs. 19:48:53 Phantom_Hoover, who said US? I aimed for world wide coverage. Anyone who is interested, from any country can join. After all, it isn't like the olympics is US only is it? 19:49:19 Vorpal, that was "us" capitalised. 19:49:26 fizzie, amb-eval sounds familiar... hm? 19:49:36 Scheme, if I'm not mistaken. 19:49:38 Vorpal: It's what SICP calls that nondeterministic backtracking thing. 19:49:39 scheme macro, backtracking? 19:49:40 right 19:49:59 I guess it could be used in related non-SICP Scheme contexts too. 19:50:05 Phantom_Hoover, oh 19:50:09 -!- MigoMipo has quit (Ping timeout: 265 seconds). 19:50:14 Phantom_Hoover, I'm sure we are not alone 19:50:25 besides you don't need to focus on the software in the TV show 19:50:26 Actually it seems to have been "ambeval", to be more accurate. 19:51:10 hm 19:51:22 Vorpal, this is never going to be a TV show. 19:51:26 EVER 19:51:50 Phantom_Hoover: Where's that AMBITION of yours now, huh!? 19:51:53 Phantom_Hoover, you could of course make each bot send out it's position if you want, that might make things a bit simpler. Especially since there seems to be more than two robots at a time 19:52:02 I was considering duels 19:52:02 Yep. 19:52:22 where keeping track of one opponent with sensors wouldn't be TOO hard 19:52:23 Duels would be good to start with, though. 19:52:27 (Oh, it's both: the evaluation procedure is "ambeval", but the prompts it prints out say "Amb-Eval".) 19:52:43 OTOH with those levels of destruction, it's basically all or nothing. 19:53:20 Phantom_Hoover, hm... that makes for fewer shows 19:53:32 It does. 19:53:47 But we're approaching this with selective practicality 19:54:15 multiple opponents changes the whole thing 19:54:23 in fact a robot might be better than a human at that 19:54:38 A CPU has no issues with keeping tracks of many things at once 19:57:54 -!- MigoMipo has joined. 20:00:36 Of course, this would be a very interesting exercise. 20:01:37 track* 20:05:49 fizzie, that robocup thing was a lot more interesting than the robowars thingy 20:06:04 well, haven't watched the complete movie 20:06:10 I get like 2 kB/s on it 20:06:18 so still not completely downloaded 20:14:52 Vorpal, stop raping my childhood memories! 20:15:43 Phantom_Hoover, ?? 20:15:53 You heard me! 20:17:21 Phantom_Hoover, yes but I didn't understand what you meant by that 20:17:36 You're MURDERING my NOSTALGIA 20:19:01 nostalgia about what? 20:22:36 BAH 20:23:34 Phantom_Hoover, robocup? 20:23:39 I said it was good? 20:23:44 or do you mean robowars? 20:23:54 how can anyone be nostalgic about that... crap 20:23:57 NOSTALGIA 20:24:01 IT ISN'T SANE 20:24:09 true 20:24:16 I WAS LIKE 5 20:24:42 Phantom_Hoover, which series? 20:24:52 I have no idea 20:24:55 they seem more insane towards the later ones 20:26:53 Phantom_Hoover, anyway I now pinned down your age that to "was around 5 at some point between 1994 and 2004" 20:27:04 Well done. 20:27:11 Phantom_Hoover, so how old are you, that way I can tell which series it was 20:27:12 I'm between 20 and 10. 20:27:21 hm 20:27:26 And I'm not sure if I was actually 5. 20:27:40 mhm 20:28:39 They have semi-recently (note more than a couple of years ago) started showing one of those shows on some Finnish TV channel; but it might not be Robot Wars exactly, since there are some others, like BattleBots. 20:28:46 Indeed. 20:28:54 And we will do it with added awesome. 20:29:12 so 2004 is unlikely, 10 years as minimum means 1990-2000 (excluding possible off by one errors due to later/earlier during the year than the current day and month) 20:29:39 Although we'll need to prevent midgets being used instead of computers. 20:29:41 (after all I'm 20, but later this year I will be 21) 20:30:27 Phantom_Hoover, why don't they make robots that can climb out of that pit btw? 20:30:30 it is not impossible 20:30:35 God only knows. 20:30:52 I always wondered why noöne had integrated a radio jammer. 20:31:00 Phantom_Hoover, probably against the rules 20:31:20 Well, perhaps the same applied to the pits? 20:31:24 hm 20:32:24 Now, more autonobot things! 20:32:33 According to http://en.wikipedia.org/wiki/Robot_Wars_%28TV_series%29#Arena_and_hazards falling into the Pit of Oblivion is instant disqualification. 20:32:38 So no use climbing up there. 20:32:49 (Besides, it sounds like a lot more added complexity for pretty little benefit.) 20:33:19 fizzie, what about folding out arms so that you hang above, and never fell into it 20:33:30 then you could climb back 20:34:20 Possibly the pit wasn't such a huge hazard. I'unno. 20:34:27 hm kevlar? 20:34:48 for added armour I mean 20:35:03 How does armour square with heat exchange? 20:35:22 I mean, you have a motor in there as well as the computer. 20:35:25 true 20:36:05 Phantom_Hoover, well didn't we agree on putting the computer off the arena? But running an AI, with no human interaction once the battle started 20:36:18 We did? 20:36:19 you still probably want a low power CPU in it 20:36:25 for some local processing 20:36:35 a high end ARM would be perfect 20:36:44 such as used in high end phones 20:36:48 How destructive do we go? 20:37:17 My god, the first series was presented by Jeremy Clarkson. 20:37:22 who? 20:37:22 Anyway. 20:37:30 anyway 20:37:30 British TV personality. 20:37:34 Indeed. 20:37:42 Phantom_Hoover, depends on what armour 20:37:53 some would act as a heatsink 20:38:07 Re destruction or heat? 20:38:12 re heat 20:38:34 Phantom_Hoover, I suggest a low-heat low-power CPU anyway if you do it locally, like one or more ARM 20:38:35 Right. 20:38:52 Phantom_Hoover, stuff like nintento DS runs on it after all. 20:38:55 Ah, well, if we're just running an ARM it's not too much of an issue. 20:38:55 and that has no fan afaik 20:39:19 Phantom_Hoover, I think a pair of high end ARMs have more than enough processing power 20:40:29 Phantom_Hoover, otherwise, since this doesn't last long, just use liquid nitrogen cooling 20:40:31 Re destruction, obviously not as insane as the show. 20:40:37 it is a one off per battle anyway 20:40:52 Phantom_Hoover, indeed, too expensive hardware at stake here 20:41:03 Yes. 20:41:15 Flippers are all right, though. 20:41:18 well yes 20:42:11 Some degree of sawing? 20:42:36 I mean, armour is pointless otherwise. 20:42:41 Phantom_Hoover, further I suggest using a composite armor 20:42:47 Phantom_Hoover, hm 20:42:55 probably 20:42:59 Phantom_Hoover, or hacking I guess 20:43:07 Yes... 20:43:25 Composite armour? Are we millionaires? 20:43:36 oh damn 20:43:43 Phantom_Hoover, titanium alloys then? 20:43:56 with steel skeleton for stiffness 20:44:23 We aren't designing tanks! 20:44:26 http://en.wikipedia.org/wiki/Titanium#Physical_properties 20:44:48 Phantom_Hoover, sure we are, titanium is light too, so you can have more of it if there are weight limits 20:44:59 The computer can be encased nicely, but everything else isn't that important. 20:45:17 Phantom_Hoover, actually *some* of those sensors might be a bit expensive 20:45:31 And anyway, we're discussing the general designs... 20:45:33 laser gyro and LIDAR primarily I think 20:45:55 you still have lots of simple switch sensors for detecting bumping into things and so on of course 20:48:19 Phantom_Hoover, anyway composite isn't that expensive. Well depends on what composite 20:51:47 OK, so robot has sensors which feed to a lightweight processor which consults a heavyweight computer off the arena and does things accordingly? 20:52:20 Phantom_Hoover, well yes the heavywight processor (if needed, it might not be) would send commands to the robot 20:53:46 Phantom_Hoover, would be free to make it at the level of "move towards that thing" or "drive on motor a", that is, you are free to put any amount of processing in the bot and any amount off it. I suggest that the CPU in the bot should not be limited in power (since other considerations will already do that) but the one off the arena should be. 20:54:00 Yes. 20:54:00 it wouldn't be so fun it someone bought a super cluster there 20:54:20 Well, it might. 20:54:37 Phantom_Hoover, might run out of space 20:54:50 You can have a supercluster, but if you still need to calculate the strategies sensibly. 20:55:42 Phantom_Hoover, I suggest simply a standard high end machine running linux off the arena. Say a dual-cpu quad-core Xeon with 8 GB RAM and a few GPGPUs attached. For each combatant 20:56:07 Yes, now can we discuss software tactics? 20:56:11 well yes 20:56:16 Phantom_Hoover, that is quite interesting 20:57:29 Phantom_Hoover, any good ideas? 20:57:54 Well... Obviously a schematic of the other robot will be useful. 20:58:10 You don't want to go charging onto a flipper, for instance. 20:58:40 Phantom_Hoover, I suggest a rouge 3D model of the opponent(s) of it marking the weapon such as flippers 20:58:44 nothing more detailed probably 20:58:49 well maybe drive system 20:58:55 Yes. 20:59:15 Marking vulnerable and dangerous points. 20:59:28 This would have to be done by a human, of course. 20:59:34 indeed 20:59:40 -!- derdon has quit (Ping timeout: 276 seconds). 20:59:56 if we allow some form of 3D system scanners (LIDAR, Radar, sonar or such), and I suggest we do, that could be matched against the 3D model loaded in advance 21:00:26 s/system scanners/scanner systems/ 21:00:33 (no idea how I managed that typo!) 21:01:34 And, as I mentioned before, being able to perform primitive diagnostics on opponents would be advantageous. 21:02:00 Phantom_Hoover, you mean position and such? 21:02:04 or more? 21:02:07 No, I mean damage. 21:02:16 I didn't say it was plausible. 21:02:25 Phantom_Hoover, well sure, that would be some interesting sensors 21:02:33 well I guess you could match against it being dented 21:03:11 But if your opponent's drive system is crippled, say, you have much more space to manoeuvre yourself. 21:03:28 You could have a human give simple hints, though. 21:03:40 Phantom_Hoover, well that would be reasonably easy to detect, if it doesn't move 21:03:48 Well, yes. 21:03:49 or moves slowly 21:04:00 Phantom_Hoover, heuristics 21:04:17 this would be good because that way you could try to pretend to be damaged 21:04:30 Phantom_Hoover, which would be quite a nice tactic 21:04:37 And thus the game becomes much more tactically complex. 21:04:43 And fun to program. 21:04:43 Phantom_Hoover, of course 21:05:24 Phantom_Hoover, anyway I wonder if it might be possible to build a drive system that can move in any direction 21:05:30 Phantom_Hoover, or nearly so 21:05:42 It is, but it'd be vulnerable. 21:05:53 how? 21:06:06 I can't see how it could be done basically 21:06:12 IIRC there are methods of doing it. 21:06:33 the best I can think of is ability to drive the left and right side wheels or tracks independently 21:06:42 which means you can turn on the spot 21:06:43 quite easily 21:06:44 http://en.wikipedia.org/wiki/Mecanum_wheel is omnidirectional 21:07:00 cleaver 21:07:07 err clever* 21:07:14 But easy to disable. 21:07:32 Compared to a wheel or track, I mean. 21:08:08 yeah 21:08:24 There's also that three-wheel design built out of what looks like the bottom bit of http://en.wikipedia.org/wiki/Omni_wheel with the wheels put like /_\ ; you can drive with two of them and the "sideways" wheels will make the third one not drag horribly. 21:10:13 fizzie, yeah, but exposed wheels can be shredded easily. 21:10:14 fizzie, still I think tracks and being able to turn on the spot is better 21:10:30 Phantom_Hoover, you can put them under your body, no need to put them at the edges 21:10:47 True. 21:11:02 Phantom_Hoover, and why not retract them and put a cover over when getting flipped? then when you flip back reveal them again 21:12:15 -!- cheater99 has quit (Read error: Connection timed out). 21:12:57 Phantom_Hoover, ruber covered metal tracks are probably better though. Rubber for added traction that is 21:13:07 and metal core for making it harder to destroy 21:13:12 Yes. 21:13:13 -!- cheater99 has joined. 21:15:56 I hate to say this, but given the amount of work that would go into building a capable autonobot, you *might* have some trouble in finding people to participate in a... how should I put it, slightly hazardous thing like that? (No matter how much you limit the destructiviness, the aim -- unless you stray pretty far -- still is to break the opponent.) 21:16:30 hm 21:17:37 How does it go with DARPA's "Grand Challenge" series of autonomous driving things; are they still doing that stuff? I haven't heard any news lately; though I guess I should just ask Google. 21:18:56 fizzie, this is the FUTURE of WARFARE. 21:19:02 We can ask the MILITARY! 21:19:48 Phantom_Hoover, in that case we want different rules 21:19:56 and composite armor 21:20:08 and probably having the crowd quite a long way away 21:20:16 I mean, missiles and such 21:21:33 There's that at-times-a-bit-hilarious paper about how having robots shooting guns is going to be more ethically good, because they can more easily in "the heat of the battle" judge who are civilians and who they shouldn't be shooting. 21:22:07 It's going to really suck for that 10 % that happen to be false negatives of the civilian-classifier algo, though. 21:22:19 indeed 21:22:41 also giving robots weapons like that? Um, you must be insane 21:23:05 Just see future predictions by Hollywood 21:24:01 Incidentally, there's quite a bit of (overview-only presentation in the upcoming video link) hardware that might be useful for an autonobot in the 2007 DARPA-winning autonocar, http://www.youtube.com/watch?v=lULl63ERek0 21:28:36 fizzie, nice but expensive 21:29:08 fizzie, btw in the mine in Kiruna they use autonomous trucks and such 21:29:20 fizzie, monitored by humans and the occasional change and such 21:29:28 but day-to-day operation is autonomous 21:30:04 fizzie, iirc they use lasers, oh and some kind of reflective things mounted on the walls at the proper height 21:30:12 so it can make sure it is not crashing into a wall 21:30:35 operations* 21:31:02 fizzie, and there is no GPS to help them down there 21:31:53 Yes, but on the other hand it's a lot more controlled environment than a random city. 21:31:57 fizzie, true 21:32:10 fizzie, they have been using them since way before those cars though 21:36:23 Meh, all this robotics talk always makes me feel like I should build something; then I remember how very non-hardwarey person I am. I think I'll stick with the bits; they're a lot more malleable. 21:38:37 There's no great shortage of hardware people. 21:38:56 fizzie, try lego, and remember to put support beams in orthogonally 21:40:21 there is really just one "secret" of lego technic: two normal "beams" with two plates between is the proper distance to mount orthogonal supports on the side 21:40:32 I could draw this in ldraw if you are interested I guess 21:41:11 ah lego has it: 21:41:13 http://cache.lego.com/2057/images/create/designschool/courses/course_1/lesson1c/und272x362brickfit.gif 21:41:28 hm cache... hope that url works for you 21:42:18 It does. 21:42:20 -!- augur has joined. 21:42:34 -!- MigoMipo_ has joined. 21:43:25 (It also reminds me of Pac-Man.) 21:43:53 fizzie, not very 21:44:13 -!- MigoMipo has quit (Ping timeout: 265 seconds). 21:44:26 Has anyone else seen the Lego Turing machine? 21:44:34 -!- MigoMipo_ has quit (Read error: Connection reset by peer). 21:45:13 Ack, now I want one. 21:45:38 Phantom_Hoover, link? 21:45:58 I seen the difference machine 21:46:01 Haven't got one, at the moment. 21:46:06 but that is not really a turing machine at all 21:46:08 It's on Youtube. 21:53:46 -!- wareya has joined. 21:56:25 -!- wareya_ has quit (Ping timeout: 255 seconds). 22:11:46 -!- sftp has quit (Ping timeout: 245 seconds). 22:15:37 -!- augur has quit (Ping timeout: 265 seconds). 22:43:53 -!- cpressey has joined. 22:52:50 -!- cpressey has quit (Quit: Leaving.). 23:03:42 Holy crap, Cubestormer is impressive 23:15:33 -!- Phantom_Hoover has quit (Ping timeout: 265 seconds). 23:22:07 ais523 is in the Crawl learndb 23:28:47 Sgeo, oh? 23:29:03 http://crawl.akrasiac.org/learndb.html#ais523 23:29:14 -!- GreaseMonkey has joined. 23:29:35 Sgeo, I wasn't aware of nethack's interface being really inconsistent 23:29:51 well, one or two small things sure but, nothing major 23:30:06 Sgeo, is crawl's interface actually inconsistent? 23:30:17 -!- fizzie has quit (Ping timeout: 240 seconds). 23:30:22 -!- fungot has quit (Read error: Operation timed out). 23:30:36 It... feels nicer than NetHack's, in that it automates tedious stuff 23:31:13 -!- FireFly has quit (Quit: swatted to death). 23:31:30 Sgeo, such as? 23:31:43 Sgeo, oh and: is it harder or easier than nethack? 23:31:59 Exploring, knowing where you stashed your stuff, note taking, resting until healed 23:32:09 Knowing what's safe to eat 23:32:42 Interlevel travel, including avoiding areas that the user marked for exclusion 23:33:16 Sgeo, knowing where you stashed stuff is easy. First level of sokoban, icebox if possible. Until right before ascending, when I move it to level 1 if feasible. 23:33:30 well the stuff I want to take with me but don't plan to carry down 23:33:48 like artifacts that I have no use for except for the score 23:34:59 price-ID is impossible 23:35:07 Sgeo, that's a pity. 23:35:24 Well, I prefer avoiding price-ID in NH 23:35:30 Since it's tedious 23:35:31 mhm 23:35:35 well yes 23:45:37 -!- Wamanuz has quit (Remote host closed the connection). 23:49:23 -!- Flonk has quit (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]). 23:49:49 -!- tombom_ has joined. 23:51:46 -!- tombom has quit (Ping timeout: 276 seconds). 23:56:12 -!- fizzie has joined.