00:00:00 it's just 00:00:04 If you can find a place to buy it from that's open. 00:00:05 awesome 00:00:10 <500 is not very low 00:00:22 umm 00:00:25 that's half the piano keyboard 00:00:37 It's low in the sense that it's "not high". 00:00:44 (which is tuned logarithmically) 00:00:54 nobody ever tunes things linearly 00:00:59 well i may agree with the incredibly low or high ones being out of the log scale 00:01:13 but 500? 00:01:14 wtf. 00:01:29 i recall making basic programs that played linear scale near there 00:01:44 it definitely didn't sound like it was linear 00:01:51 and i didn't know shit about music then 00:02:05 They just gave people a comparison tone, and asked them to adjust another one to be half as high. 00:02:27 Non-musicians, I would guess. 00:03:18 well maybe musicians who don't know shit about frequencies 00:03:19 dunno 00:04:06 but wtf, "half as high", i'm not sure that'd make any sense to me if i didn't know that meant an octave. 00:04:56 i'd probably think it's impossible because you can get as low as you want 00:05:10 Alternatively people were asked to divide a frequency range to "equally large" regions. 00:05:21 i might be content with approximating half of one interval 00:05:24 yeah oka 00:05:25 y 00:05:39 It's so very difficult to live outside the campus area; all papers are "buy this article for $19 since you don't have a subscription". 00:05:50 Should set up the proxying stuff some day. 00:06:28 fizzie, heh 00:07:28 There seems to be an ICASSP paper from 1999 whose abstract says they "investicate the often made remark that there are two regions to the Mel scale, the first region (" but then it cuts off and the full text is only available for a fee. 00:07:56 :) 00:08:04 well that's prolly the gist of it. 00:08:40 "We show that there is no evidence, based on the experimental data points, that there are two qualitatively different regions or that the lower region is linear and upper region logarithmetic." 00:08:47 (Started a browser with X forwarding.) 00:08:58 Heh, it's a very widely distributed rumour, then. 00:09:57 nothing like that fits my intuition the least bit, and i know nothing about sound; would say that's a pretty weird rumour. 00:10:14 Apparently the points aren't exactly logarithmic either, and it's mostly just a matter of how do you want to approximate it. 00:11:08 how did you get the rest of it? 00:12:24 I started a browser at one of the computers inside the campus, used their IEEEXplore submission to get the full text. 00:12:55 that's some serious hacking right there 00:13:30 Basically seems like they've taken the data points from the original Mel paper (from the 1940s) and noticed that since there's a whole pile of curves that fit the points, you can't really say much more than that it's not linear, it's not logarithmic, and it can be numerically approximated with various functions. 00:14:24 This is a lot of discussion to be having for slide #3, since the whole presentation has something like 17 of 'em. :p 00:14:48 yeah i'm sure the presentation was like this too 00:15:01 really i'd forgotten context loooong ago 00:15:22 Actually it went really fast; we were running a bit late at that point, so I skipped most of the boring parts. 00:15:36 how does this 'mel' scale correlate with the greenwood function btw? 00:16:40 Not sure; that's of course the physiological justification for it. That "position in the ear" thing is a logarithmic approximation too, right? 00:18:07 yeah, greenwood is log (1 + f / a) / k, just with different constants 00:19:22 What happens after the cochlea is probably far less understood than the "just a simple matter of physics" part. 00:21:03 There's quite a lot of speech recognition research that's been motivated by "it seems as if humans do it sort of like this, a bit, maybe"; it's a bit of a buzzword nowadays. 00:21:45 Personally I think half of those people just want any convenient excuse to go stick electrodes into brains. 00:22:19 heh, it's a convenient idea that by just mapping the raw data onto what's been measured from sensory organs you'll get something profound and/or easily analyzable 00:23:34 'night 00:23:46 -!- Mony has quit ("Join the Damnation now !"). 00:28:56 hmm. now i kinda wanna go buy some coke. but i'd have to do it. 00:28:59 fizzie: fungot 00:29:29 *please, fungot 00:30:28 oklo: the drink or the drug 00:30:58 this time i'll leave it for guesswork 00:31:06 oklo: i assume the latter, for "do it" 00:31:21 ehird: assume what you wish, big boy 00:31:29 oklo: i see. 00:31:36 (for reference, your nick is longer than mine) 00:31:44 -!- oklo has changed nick to oklopol. 00:31:55 oklopol: rename yourself to ok 00:31:57 and god it smells like pee here 00:31:57 or oko 00:32:12 the dog pees everywhere, and i don't wanna sweep it up all the time 00:32:29 that's another thing i'd have to do for it to happen 00:33:51 -!- fungot has joined. 00:33:59 There's the 'got. 00:34:19 first of all, i need to check if you've tampered with ^bool 00:34:28 so don't mind my small-scale flood here. 00:34:29 ^bool 00:34:29 ^bool 00:34:29 ^bool 00:34:29 ^bool 00:34:29 No. 00:34:29 Yes. 00:34:29 No. 00:34:29 No. 00:34:31 ^bool 00:34:31 No. 00:34:33 ^bool 00:34:33 Yes. 00:34:35 ^bool 00:34:35 Yes. 00:34:42 I was just at Tampere, it's very likely that I've been tampering. 00:34:44 okay, looks about right. 00:34:50 should i go buy some cokers? 00:34:51 ^bool 00:34:51 Yes. 00:34:57 he said yes! 00:35:01 * oklopol is gonna go! 00:35:05 happy time \o/ 00:35:08 wooohhoo 00:35:09 All systems are go. 00:35:10 -> !!! 00:35:14 oklopol: which kind of coke, i ask again 00:35:16 XD 00:35:45 ;)))))) 00:35:47 -> 00:36:17 due to his unresponsiveness I shall assume that oklopol will return on cocaine. 00:36:19 XD 00:36:22 hmph, the dog was sleeping so sweetly, and now i'm gonna leave and she'll be waiting for me face squeezed against the wall :< 00:36:31 or the coke fuel 00:36:33 he might be buying that 00:36:35 's possible. 00:37:12 that does take some of the fun away, i have a lot of compassion for lesser creatures 00:37:37 LIKE DOGS, BOTS, POLITICIANS AND WOMEN 00:37:46 -> 00:37:46 lulz 00:39:34 awww oklopol no marries? :( 00:41:05 you can have 0.4 marriage, he said. 00:41:10 :D 00:41:35 I is the sleep now; night. 00:55:38 night 02:28:29 i think i should be a sleep too, now 02:28:30 -> 02:33:10 oklopol can't marry 02:33:17 Because GAYS CAN'T MARRY 02:33:19 Zing! 03:05:49 -!- Corun has quit ("This computer has gone to sleep"). 03:28:46 -!- Slereah_ has quit (Read error: 60 (Operation timed out)). 03:29:00 -!- Slereah has joined. 04:43:05 -!- Slereah has quit (Read error: 60 (Operation timed out)). 04:45:28 -!- ab5tract has quit. 05:59:40 -!- Sgeo has quit (Remote closed the connection). 07:11:13 -!- Slereah has joined. 07:15:55 theres a wonderful proof by dewdney of the nonexistance of a TM that solves the halting problem 07:15:57 its quite nice 07:15:58 :o 07:17:07 Lynx plz 07:17:13 Turing's proof is pretty meh. 07:17:25 what was turings proof? 07:19:17 Lemme check, I don't remember too good 07:19:29 First of all, it's not about the halting problem 07:19:39 Because his original specs was not halting. 07:20:27 He had two types of symbols. Symbols of the first kind (usually numbers), that are part of the final result 07:20:44 And symbols of the second kind, which are... not. They're markers and such. 07:20:57 ok well heres dewneys Devilmachine example 07:21:22 A circular machine writes down a finite amount of symbols of the first kind, a circle-free machine does not. 07:21:37 The halting problem is hence how to know if the machine is circular. 07:21:42 suppose that we have a mean H, that can decide whether or not a machine T will halt on some input i 07:21:57 Are you going to feed H in H? 07:22:00 you do this by giving H a description dT of the machine T 07:22:04 With H as input? 07:22:15 (no, not H, another machine derived from it, hold on :P) 07:22:45 so if T halts on i, H goes into 1 and halts 07:22:53 if T doesn't halt on i, H goes into 0 and halts 07:23:31 Yeah yeah, I know the oracle machine 07:23:36 Now make with the proof :o 07:23:50 now suppose we derive a second machine, DM, build from H 07:24:12 such that it has one input, a machine description, which becomes dT and i 07:24:27 for the version of H that is wrapped up inside of DM 07:25:01 and now suppose we also change 1 so that when DM transitions to 1, it unconditionally transitions back to 1, looping around on forever 07:25:13 so does DM halt on dDM or not? 07:25:15 assume it does 07:25:40 if DM halts on dDM, then it'll go into state one, and thus it'll loop forever 07:25:58 I think I lost you at H that is wrapped up inside of DM 07:26:01 ok 07:26:08 you derive a machine from H 07:26:25 How? 07:26:58 by adding a "copier" onto its inputs, so that the copier takes dT and duplicates it, putting it as the dT and i of the inner version of H 07:27:37 its just a machine that is exactly like H, only dT and i are always identical 07:27:52 Oh. 07:27:54 ok? 07:28:02 and you also add an infinite loop to state 1 07:28:11 Wait, what is the dT in question? 07:28:21 What machine does it describe 07:28:29 any machine you want 07:28:36 psygnisfive: you do this by giving H a description dT of the machine T 07:28:39 that's what H does 07:28:46 given some machine description, and an input i 07:28:53 it determines if the described machine halts on i 07:28:54 Why are you talking to yourself 07:29:02 im not, im QUOTINg myself 07:29:12 anyway 07:29:15 What IRC program are you using? 07:29:23 limechat, why 07:29:35 anyway 07:29:38 do you understand?? 07:29:43 'cause they usually have 07:29:44 Yeah. 07:29:47 ok 07:29:47 so 07:30:10 then you give dDM to H 07:30:13 as both input and output 07:30:21 asking, will DM halt on its own description? 07:30:30 if H says yes then this is what happens in DM: 07:30:54 DM takes dDM as input, and because it has an inner copy of H, the inner copy of H should ALSO say yes 07:31:01 which means that it it goes into state 1 07:31:20 but DM has an infinite loop on state 1, so DM cannot halt on its own input, if H says it will 07:31:35 I think that's pretty much Turing's proof. 07:31:40 so maybe H will say no, DM wont halt on its own input 07:32:00 if thats the case H will go into 0, meaning so will DM, and 0 is a halting state for DM 07:32:03 so DM WILL halt 07:32:20 so either way, H is wrong, and this machine, DM is possible if H is possible 07:32:29 thus since DM is not possible, H is not possible 07:32:45 its a beautiful proof 07:33:26 But I think it's the proof of Turing, apart from the letters. 07:33:30 tho it helps to hav diagrams :) 07:33:35 it might be turings proof 07:33:36 i dont know 07:33:53 all i know is, in class the professor had multicolored chalk 07:34:03 so the diagram for H was clean and white 07:34:12 It's a machine D, that when given the standard description of a machine M, test if M is circular, and mark the SD with u, otherwise with S. 07:34:29 and the diagram for DM was purple and orange and green and there were all sorts of swirlies here and there representing arbitrary input and stuff 07:34:42 so it looked psychedelic and demonic and appropriate for a machine named the Devil Machine 07:34:44 i loved it 07:34:50 pix 07:34:57 no pix :( 07:35:45 Hm. 07:35:51 I think I see where theproof is different 07:36:08 Turing applied it directly to the problem of the extension of the computable numbers. 07:43:25 Hm. 07:43:27 Teabagging. 07:43:43 <3it 07:45:18 I BET YOU DO, FAGGOT 07:45:30 Also : http://en.wikipedia.org/wiki/Beard_Liberation_Front 07:45:37 I'm afraid I must declare holy war on you 07:56:55 oh i love beards mate 07:56:59 just not on the guys i fuck 07:57:14 Do you like disembodied beards? 07:57:24 Like Le Chuck's beard in Monkey Island 2 07:57:24 yes, very much so 07:58:42 i have to say, im not impressed by the 2001 beard of the year 07:58:51 What if I sort of shaved my balls, except on the bottom 07:58:56 So that they have a little beard. 07:58:58 lol 07:58:59 hahahahha 07:59:17 i'd rip the hairs out all at once 07:59:17 Then I can put on a face with a marker 07:59:43 I would slap you with my penis. 07:59:59 its an interesting thing that people wish facial hair have no chance in american politics on a large scal 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 08:00:01 e 08:00:33 What about ABRAHAM LINCOLN 08:00:44 Or FRANKLIN ROOSEVELT 08:00:48 modern american politics i mean. 08:00:55 i think roosevelt was the last bearded president 08:01:00 well 08:01:03 TEDDY was 08:01:07 franklin was clean shaven 08:01:17 Didn't he have a moustache. 08:01:21 no 08:01:23 that was teddy 08:01:25 I'm pretty sure Obama has a ball beard though. 08:01:43 oh no taft had a mustache 08:02:27 he was the last 08:02:28 taft 08:04:20 There's too much presidents with the same name. 08:04:37 theres only two pairs of same-named presidents in american history :P 08:04:42 Good thing Hillary wasn't elected 08:04:49 well 08:04:49 three 08:04:50 That's too much for me 08:04:52 bush included 08:05:00 Who's the third? 08:05:18 the bushes, the roosevelts, and the .. other guys from the beginning 08:05:20 adamses 08:05:23 i think 08:05:32 Oh. 08:05:35 Wait... 08:05:41 John Adams and... Quincy Adams? 08:05:53 John and John Quincy 08:05:54 yeah. 08:05:56 I mostly know my US Presidents from the Animaniacs song. 08:06:21 interestingly 08:06:27 John -> John Quincy 08:06:32 George -> George W 08:06:40 http://fr.youtube.com/watch?v=Vvy0wRLD5s8 08:06:43 both father-son lineages 08:07:53 andrew jackson had some crazy clothes 08:08:05 http://en.wikipedia.org/wiki/Image:Andrew_jackson_head.gif 08:08:07 i mean what is that? 08:08:28 He's a wizard 08:08:35 :o 08:24:57 -!- ineiros has joined. 09:23:23 -!- Mony has joined. 09:23:35 plop 09:25:11 'sup. 10:53:22 -!- habnabit has changed nick to JUMP_ABSOLUTE. 10:59:17 -!- KingOfKarlsruhe has joined. 11:01:53 -!- Slereah_ has joined. 11:20:44 -!- Slereah has quit (Read error: 110 (Connection timed out)). 12:06:58 -!- puzzlet_ has quit (Remote closed the connection). 12:07:05 -!- puzzlet has joined. 12:24:20 -!- lifthrasiir has quit (Read error: 110 (Connection timed out)). 13:03:23 -!- KingOfKarlsruhe has quit (Remote closed the connection). 13:16:07 -!- puzzlet has quit (Read error: 104 (Connection reset by peer)). 13:17:13 -!- puzzlet has joined. 13:48:01 i don't know if psygnisfive's is turing's proof, but that's the common one, except there's usually less wrapping and descriptions. 13:51:46 http://clisp.cvs.sourceforge.net/*checkout*/clisp/clisp/doc/Why-CLISP-is-under-GPL I love the "Is that what you want?" line. rms is and always has been a fear-mongering person who cannot handle people with other views on commercial software. His position on the linking is totally ridiculous ... he could probably claim that every program on my system must be licensed under the GPL with that. 13:52:05 I wish I was cynical enough to call his sudden change of heart mind control. :-p 13:53:06 i don't understand how you people can keep track of pop culture so well, i have no idea what rms means 13:53:16 richard m stallman, creator of the gnu project and gpl. 13:53:24 Richard Stallman is hardly pop culture... 13:53:26 hobo. 13:53:28 (moron.) 13:53:36 note: last two may be personal opinion 13:53:51 oh 13:54:05 well i know stallman, i'm not *that* much out of the looper 13:54:11 you should be 13:54:13 quick, forget stuff 13:54:19 :) 13:54:42 well as ais523 said, as long as it's not pop pop culture, it's okay, it's just not not okay if i *don't* know. 13:55:08 wat 13:55:17 s/it's just/and it's 13:55:33 what's that kind of a mistake called? 13:56:05 i use terms that show something implicitly, and the implicit parts have inconsistency 13:56:21 !(pop pop culture) => 13:56:21 OK 13:56:21 & ((not know) => OK) 13:56:22 right? 13:56:59 (OK ((not (pop pop culture)) => (not know)) 13:57:06 hmph 13:57:12 (OK ((not (pop pop culture)) => (not know))) 13:57:26 (OK (not (pop pop culture)) 13:57:36 (OK (not know (not (pop pop culture)))) 13:57:37 wait, what 13:57:38 right? 13:57:42 that's what you said 13:57:48 no no that's not right, i was just copypasting yours :P 13:57:53 let's try predicate logic. 13:58:48 (forall X)(okay(pop_pop_culture(X) ^ know(X))) 13:59:59 where ^ equals AND, or XOR? 14:00:09 "as long as it's not pop pop culture, it's okay, and it's not not okay if i *don't* know." (forall X)(okay(pop_pop_culture(X) ^ know(X))), then emphasis on (not (not (okay(pop ... know(X))))) 14:00:20 ais523: and. 14:00:25 ais523: and, it's predicate logic 14:00:32 wait 14:00:37 oklopol: that's wrong 14:00:38 that says: 14:00:44 well no wonder that looks wrong :) 14:00:45 it's okay if (it's pop pop culture and I know it) 14:00:48 it has an extra not 14:00:53 ah, okay 14:01:00 soo you need an extra not there 14:01:01 yeah my predicate logic thing has a copypaste error 14:01:09 (forall X)(okay(not(pop_pop_culture(X)) ^ know(X))) 14:01:18 by pop pop culture 14:01:23 i meant the more popular pop culture 14:01:38 i then quickly forgot what that meant and started getting confused 14:02:07 now let us start over having learned from these past mistakes 14:03:52 "as long as it's not pop pop culture, it's okay, and it's not not okay if i *don't* know." (forall X)(okay((((not pop_culture(X)) ^ (pop_pop_culture(X))) ^ (know(X))))) 14:04:30 the second sentence is just gravy, and the first and in the predicate logic is linguistically a "but" 14:04:42 LOL 14:04:56 and once again i have the two pop cultures the wrong way around :P 14:05:10 i conjecture i cannot achieve consistensy in terminology 14:05:22 * ehird bear-poops 14:06:06 but yeah if you swap those, it should be fine 14:26:54 -!- KingOfKarlsruhe has joined. 14:31:43 -!- Corun has joined. 14:40:15 -!- sebbu2 has joined. 14:44:31 -!- Judofyr has quit. 14:52:41 -!- kar8nga has joined. 14:58:15 -!- ehird has left (?). 14:58:15 -!- ehird has joined. 15:00:15 -!- sebbu has quit (Read error: 110 (Connection timed out)). 15:07:07 -!- Judofyr has joined. 15:28:44 -!- puzzlet_ has joined. 15:29:33 -!- puzzlet has quit (Read error: 60 (Operation timed out)). 15:51:33 -!- Asztal has joined. 16:12:33 hm 16:12:40 ais523, hi 16:12:47 hi 16:12:52 ais523, I'm currently working on conditional fingerprints 16:12:59 what do you mean by that 16:13:05 fingerprints which are included in some builds not others 16:13:06 like IFFI 16:13:09 ? 16:13:14 basically: A fancy name for adding an #if around the fingerprint 16:13:19 so it can depend on checks in cmake 16:13:22 for example 16:13:30 if we don't have ncurses skip TERM 16:13:31 or such 16:14:11 also my fingerprint managing scripts are looking about as messy as ick's old main() now. 16:14:20 he 16:14:21 *heh 16:14:30 including a ad-hoc "generate fingerprint list for man page"-mode 16:14:32 an* 16:15:02 I should probably split out the common code (parse *.spec files) into a separate file 16:15:08 and then source it 16:28:46 ais523, hm a question that I suspect you may know the answer to: 16:29:01 go on 16:29:03 -!- kar8nga has left (?). 16:29:06 what if you have a C file that is empty? 16:29:14 apart from possible whitespace 16:29:30 is that legal? 16:29:36 I don't think so 16:29:36 according to the C standard 16:29:42 all files have to have at least one declaration in 16:30:01 ais523, ok, then what if the C file isn't empty, but everything is in an #if ... #endif that evaluate to false 16:30:09 logically it would be empty after the pre-processor then 16:30:19 is that file valid or not? 16:30:19 yes, it would 16:30:22 so I think not 16:30:29 write int foo; and it becomes legal 16:30:49 ais523, gcc complains for truly empty files but not ones that are empty due to everything being iffed out 16:31:21 ah, ok 16:31:29 an extension, but an obvious one 16:31:33 ah I know a better way to solve it, include the global header before, it contains some typedefs 16:31:40 I guess that would be enough 16:31:56 or? 16:32:09 not sure 16:32:14 I can't remember the exact wording 16:32:21 comp.lang.c are good for this sort of thing, I find 16:32:46 ais523, hm, I'll take a look at the C99 spec. I guess. 16:36:58 -!- KingOfKarlsruhe has quit (Remote closed the connection). 16:40:55 ais523, hm from now on I fear that libick_ecto_b98.a won't compile without adding a few HAVE_*. At least HAVE_random, HAVE_srandom and HAVE_clock_gettime 16:41:06 ah, ok 16:41:15 so I should probably try to get the script integrated with your build system 16:41:17 hmm... maybe we can write fallback libraries 16:41:20 I may do that later 16:41:31 ais523, hm? 16:41:43 if srandom isn't there 16:41:46 provide our own version 16:41:47 etc 16:41:50 well I fall back to srand() 16:41:54 ah, and ok 16:41:55 makes sense 16:42:00 and for clock_gettime() I fall back to gettimeofday() 16:42:04 and yes 16:42:32 Since I try to be POSIX.1-2001 + _POSIX_MAPPED_FILES + strdup 16:42:46 strdup and srandom/random are XSI 16:43:21 (which is a superset of POSIX, XSI is required for a system to be Unix however, see SUSv3) 16:43:43 oh and all those except srandom/random are mandatory in POSIX.1-2008 16:43:43 ok 16:44:11 but that will be at least a few years before I start depending on the 2008 edition 16:47:33 btw, IFFI will not need an update for the conditional fingerprint bit due to both versions 1.2 and 1.3 of the fingerprint spec format are supported atm, since 1.3 only added an optional line (%condition:...) 16:47:46 ok 16:49:55 -!- Judofyr_ has joined. 16:49:55 -!- Judofyr has quit (Read error: 104 (Connection reset by peer)). 16:56:13 ais523, in English, if you put a whole sentence within parentheses do you put the . before or after the ) ? 16:56:30 such as (This is a whole sentence). or (This is a whole sentence.) 16:56:35 inside 16:56:45 well, the . is inside the parens 16:56:45 ah right 16:57:05 the basic rule is that the sentence needs to be punctuated correctly if you remove everything inside the parens 16:57:43 ok 16:58:10 hrrm, what bash version do you have? (echo $BASH_VERSION) 16:58:57