00:00:31 <lament> congratulations, you're warmer than the surroundings
00:00:47 <lament> that means you're probably alive (or died only recently)
00:01:57 <oklopol> also the cat can smell your love for it, and it gets a warm fuzzy feeling outta that
00:16:44 -!- puzzlet has quit (Remote closed the connection).
00:16:48 -!- puzzlet has joined.
00:35:23 -!- nooga has quit ("Lost terminal").
01:32:15 -!- sebbu has quit ("@+").
01:59:06 * oklopol needs to wake up in about 3 hours :<
02:17:41 -!- Asztal has quit (Read error: 104 (Connection reset by peer)).
02:37:03 -!- puzzlet_ has joined.
02:43:41 -!- Corun has quit ("This computer has gone to sleep").
02:48:44 -!- puzzlet has quit (Read error: 110 (Connection timed out)).
03:43:11 -!- oerjan has quit ("leaving").
03:48:26 -!- puzzlet_ has quit (kornbluth.freenode.net irc.freenode.net).
03:48:50 -!- puzzlet_ has joined.
06:20:39 -!- habnabit has joined.
06:21:37 <habnabit> Hi guys! I imagine you don't see much python code, so I decided to fix that.
06:21:39 <habnabit> http://paste.pocoo.org/show/91848/
06:23:00 <habnabit> Who says you can't do functional programming in python?
06:38:48 -!- habnabit has quit (Remote closed the connection).
06:39:09 -!- habnabit has joined.
06:43:47 <lament> habnabit: please tell me it's not hand-written
06:44:33 <fizzie> Probably the same people who say you can't do functional programming in Perl, yet perl -e 'print &{&{sub { my $f = shift; return &$f($f); }}(sub { my $f = shift; return sub { my $n = shift; return 1 if $n < 2; return &{&$f($f)}($n-1) + &{&$f($f)}($n-2); }; })}(11), "\n";' is a perfectly fine and very functional recursive fibonacci.
06:44:33 <habnabit> lament, it is 100% hand-written.
06:45:04 <lament> hm... what is it supposed to do?
06:45:10 <lament> it just terminated without output
06:45:23 <habnabit> You have to add 'main(some_port_number)' at the end if you want to run it.
06:45:26 <lament> oh, duh, it's a function definition
06:45:46 <habnabit> telnet works as a client. It uses CRLF as a delimiter.
06:46:15 <habnabit> It handles SIGINT and SIGTERM gracefully, too.
06:47:55 <habnabit> So yeah. I probably am crazy, but I love writing obsfucated python.
06:51:32 -!- bsmntbombdood has changed nick to xor.
06:52:27 -!- xor has changed nick to bsmntbombdood.
06:55:19 <pikhq> For my CS class, I've been asked to do merge sort and quick sort without recursion.
06:55:28 <pikhq> So, I have implemented a full call stack in C.
06:55:38 <pikhq> (instead of the somewhat crazy way they wanted it done)
06:57:12 <pikhq> No, though that is tempting.
06:57:21 <pikhq> Are you familiar with the unary && operator?
06:58:07 <pikhq> && returns the address of a goto label.
06:59:25 <pikhq> Stack of that and the rest of the state of my code, and a few gotos.
07:00:32 <GregorR> pikhq: I got a new server for Codu, Plof Hg is now accessible via https :P
07:13:29 <fizzie> That's not "C", though -- that's GCC.
07:17:00 <pikhq> fizzie: It's GNU C.
07:17:25 <pikhq> Fortunately for me, this class uses GNU C++... Which means I can do this crazy shit and get away with it.
07:39:45 -!- Sgeo has quit ("Ex-Chat").
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:18:50 -!- jix has joined.
08:18:57 -!- jix has quit (Remote closed the connection).
08:19:04 -!- jix has joined.
08:19:09 -!- jix has quit (Remote closed the connection).
08:44:30 -!- fungot has quit (Read error: 145 (Connection timed out)).
08:44:46 -!- fizzie has quit (Read error: 145 (Connection timed out)).
09:11:07 -!- olsner has quit ("Leaving").
09:32:09 <oklopol> oh, seems i'm scrolled up.
09:32:18 <oklopol> that makes less sense in whatever context i happen to be saying it in.
09:33:01 <oklopol> habnabit: i have an irc bot that runs brainfuck written in that style
09:33:07 <oklopol> although it isn't obfuscated
09:34:32 <oklopol> http://www.vjn.fi/oklopol/python.txt though that's just the interp, but the irc stuff is trivial ofc
09:34:46 <oklopol> not sure where i have the actual botter
09:41:03 <habnabit> Though I cheat a little and use a dynamically-created dict subclass as a namespace instead of abusing function arguments.
09:41:26 <oklopol> yes that would be less verbose
10:18:01 <pikhq> oklopol: Obfuscation is easy.
10:18:22 <pikhq> And using goto in a low-level CS course for good reason is just silly.
10:38:07 <oklopol> pikhq: well i think trivial is a special case of easy, but i agree, it's easy.
10:39:05 <oklopol> and yeah, goto is often silly
10:45:24 <pikhq> What's silly is using goto to implement a call stack.
10:45:46 <oklopol> yes i read your idea, was pretty great imo
10:46:31 <pikhq> Unfortunately, the code in question works on my desktop computer, not the CS labs' computers...
10:46:44 * pikhq shakes his fist at GCC 3
10:48:07 <oklopol> hmm. now you made me want to do something.
10:49:18 <oklopol> an exercise in AOCP made me think it might be pretty cool to make like an esoteric programming book, would be pretty esoteric
10:49:42 <oklopol> like, introduce the best languages, and techniques for different paradigms
10:50:11 <pikhq> That would be pretty cool.
10:50:18 <oklopol> and perhaps some more formal treatment too, P'' and TM's and shit
10:50:31 <oklopol> i was thinking of a three-part structure
10:50:58 <oklopol> one part that's about compiling languages to others, that is, making layers of abstractions on languages
10:51:18 <oklopol> characterize the overhead of paradigm changes a bit, and stuff like that
10:51:52 <oklopol> then, something about programming directly in esolangs, visualizing things to yourself, and golfing, naturally
10:52:00 -!- KingOfKarlsruhe has joined.
10:52:17 <oklopol> i would have tons to say about these subjects, but i don't know if i'm structured enough to make an actual book.
10:52:30 <oklopol> and i'm a bit too lazy to try... i think, maybe i could try to try.
10:53:40 <oklopol> for a third part i was thinking the more formal stuff, proving esoprograms right and complexities, more formal treatment about the overhead of compilation between paradigms
10:55:16 <oklopol> and of course, tons of exercises, would definitely be the most fun set of exercises ever, if done well (non-biased objective opinion)
10:57:10 <oklopol> how's it hanging heathens?
11:04:49 <oklopol> i don't even know tbh. i'm so tired and hungry i'm talking in metaphors and mirages.
11:05:58 <oklopol> i slept about 2.5 hours last night
11:07:36 <pikhq> I'm thinking at the rate I'm going, I'm going to completely omit sleep today.
11:09:27 <oklopol> i pretty much have to read, or i won't reach my daily reading quota
11:10:14 * oklopol is slowly turning into a bibliophiliac, or whatever the term is
11:13:36 <oklopol> dream theater instrumentals = perfection
11:27:06 -!- Slereah has joined.
11:31:51 <Slereah> Me too, 'cause it's actually in 10 monts.
11:32:36 <oklopol> but my age will be a composite number for ages after that :<
11:32:36 <Slereah> BUT MINE WAS A SHORTER AMOUNT OF TIME BEFORE
11:33:12 <Slereah> Also, I have to actually program something useful.
11:33:52 <Slereah> Like I morally should not do this D:
11:34:11 <Slereah> I have to program a simulation of a thousand particles.
11:34:31 <oklopol> well can you at least write it in an unlanguage?
11:35:01 <oklopol> pikhq is making just the right tool for you then
11:35:43 <Slereah> A particle automaton or something?
11:36:07 <oklopol> his making c recursion prettier
11:36:37 <Slereah> Well, it's to be done in regular c, me think!
11:36:48 <Slereah> Plus, I won't use recursion I think.
11:37:09 <Slereah> One over all time and one over all particles.
11:37:12 <oklopol> you have like gravities and shit?
11:37:29 <Slereah> Nah. It's the simulation of a gas.
11:37:44 <Slereah> 1000 classical particles, with three collision probabilities.
11:37:59 <oklopol> what does that latter thing mean
11:38:00 <Slereah> (elastic collision, one with losses and one with gain)
11:38:31 <Slereah> It means that after the collision, the particle has more kinetic energy
11:38:49 <Slereah> It's a quantum thingamagig.
11:39:10 <oklopol> rright, yeah, i wouldn't know anything about that
11:39:24 <oklopol> what are you aiming for with this?
11:39:32 <oklopol> something revolutionary i bet
11:39:36 <Slereah> Getting a good grade, mostly.
11:39:43 <oklopol> i've only ever written programs with *one* loop
11:39:53 <oklopol> this is a whole new dimension for me
11:40:43 <oklopol> so no gravity? you can boost this quite a lot with hashing then
11:41:01 <pikhq> Slereah: I was asked to do merge sort and quick sort without recursion.
11:41:06 <Slereah> The collisions aren't really done though, you know, a collision.
11:41:16 <pikhq> So, I implemented a call stack. Hooray.
11:41:31 <Slereah> It's an application of the class on random number generators.
11:41:47 <Slereah> Of course, it's a Boltzmann distribution, so it's a bit harder D:
11:42:34 <oklopol> Slereah: right, i have no idea about any of this
11:47:22 <AnMaster> Slereah, are you looking for a good RNG?
11:48:17 <AnMaster> http://en.wikipedia.org/wiki/Blum_Blum_Shub is fairly secure iirc though slow
11:48:51 <Slereah> No need, we use the generator they gave us in class.
11:49:31 <Slereah> It's a mix of the (a * n) + b thingamagig and the registry moving one.
11:49:42 <Slereah> "Gnrateur raliste" we call it.
11:50:07 <AnMaster> Slereah, I'm not really an expert on that stuff
11:50:28 <Slereah> http://en.wikipedia.org/wiki/Linear_congruential_generator
11:50:49 <Slereah> And... that one, I think http://en.wikipedia.org/wiki/Linear_feedback_shift_register
11:52:36 <Slereah> That's only to generate an equiprobable number between 0 and 1 though
11:52:45 <Slereah> The hard part is to do the distribution.
11:53:42 <AnMaster> http://en.wikipedia.org/wiki/Yarrow_algorithm looks nice
11:53:57 <AnMaster> Slereah, what are you going to use ifor?
11:54:39 <Slereah> Well, in a box of particles, the speed of those particles at thermal equilibrium is defined by a statistics.
11:55:02 <Slereah> The simplest model is the Maxwell-Boltzmann statistic.
11:55:15 <AnMaster> http://en.wikipedia.org/wiki/Mersenne_twister
11:55:18 <Slereah> Nah, as said, we use what we get in class.
11:55:21 <AnMaster> was developed for simulations iirc
11:55:32 <Slereah> Apparently it's what they use at work!
11:55:50 <Slereah> That's why I feel dirty, AnMaster
11:56:00 <Slereah> I have to do reasonable things D:
12:31:17 -!- Corun has joined.
13:02:53 -!- ehird has left (?).
13:16:12 -!- ehird has joined.
13:50:23 -!- oerjan has joined.
13:51:08 -!- KingOfKarlsruhe has quit (Remote closed the connection).
14:05:35 -!- Slereah_ has joined.
14:12:05 <ehird> an algo that loops through an array twice is O(n), right?
14:12:09 <ehird> since O(n*2) = O(n)??
14:12:49 <Slereah_> Since it refers to the degree of the polynom.
14:13:01 <Slereah_> I think. I mostly use the O that means pretty much the opposite.
14:14:00 <Slereah_> (O(x) in science means ax + bx + cx^3 + ...)
14:18:01 <oerjan> well, one is the limit as n -> infinity, the other as x -> 0
14:18:19 <Slereah_> Yeah, that's the rigorous definition.
14:18:29 <Slereah_> When you do physics, though, it's pretty much that.
14:18:40 <Slereah_> 'cause it usually pops up from infinite series.
14:18:44 <oerjan> and a function can be O(x) without being analytic
14:19:17 <Slereah_> It can be O anything, technically.
14:20:21 <Slereah_> But what amuses me is, in computer science, a big O is bad.
14:23:05 -!- Slereah has quit (Read error: 110 (Connection timed out)).
14:30:02 <ais523> ehird: for i=1..n: for j=1..n: do stuff is O(n^2); for i=1..n: do stuff; for j=1..b: do stuff; is O(n) (=O(2n)).
14:48:58 <oklopol> 16:13… Slereah_: (O(x) in science means ax + bx² + cx^3 + ...) <<< i don't get it, explain
14:49:23 <ehird> http://qntm.org/?gay graph marriages
15:09:44 <oklopol> but yeah the idea of graph marriages is nice, but we definitely need directed edges, and edge weights
15:10:08 <oklopol> i mean, i'm not gonna marry psygnisfive no matter how many times he asks, but i guess i could marry him a *little*, that wouldn't be so bad.
15:11:05 <oklopol> i guess there's already facebook for that kinda thing
15:13:33 <ehird> marry him a little
15:14:11 <ehird> "oklopol... will you marry me 0.5714?"
15:14:38 <oerjan> "make that 0.4836 and we have a deal"
15:14:58 <oklopol> "my heart already belongs 0.683 to someone else, but i guess i could give you like 0.249"
15:15:54 * oerjan was hoping for some good old haggling here
15:16:38 <oklopol> engagement wouldn't be needed either, you could just marry people gradually, over time
15:16:51 <oklopol> maybe there could be functions over time attached on the edges!!
15:17:06 <oklopol> so you could get married over the course of two years
15:18:38 <ehird> i'm writing botte discuss.
15:19:25 <oerjan> to wryte a botte is a moste eloquente thinge
15:19:28 -!- fizziet has joined.
15:19:39 <oklopol> fizziet: okay *that's* feminine
15:19:40 <fizziet> I get so utterly bored in trains.
15:20:02 <ehird> that was some disjointed conversationn
15:20:17 <fizziet> Hey, my normal client is not here.
15:20:24 <oerjan> i think i'll skip the obvious joke
15:20:28 <oklopol> yeah, so you don't even have an excuse!
15:20:55 <oerjan> fizziet: although you are perfectly poised for making shakespeare jokes
15:21:21 <oerjan> although really we would need hiato here...
15:21:59 <oerjan> and this would nicely converge the two last conversations
15:22:00 <oklopol> fizziet: have you considered sleeping?
15:22:35 <oerjan> because it ends with -o
15:23:45 -!- oklopol has changed nick to okleo.
15:24:20 <okleo> o fizziette, how beautifully is thine facette on this day so brighte!
15:24:29 <okleo> i could so write shakespeare
15:25:02 <oerjan> oh fungot wherefore art thou fungot?
15:25:16 <ais523> and where did thututbot go, I wonder?
15:25:17 -!- fizzie has joined.
15:25:25 <ehird> it is just a good thing i am writing botte
15:25:27 <ais523> oh, ehird rebooted rutian
15:25:36 <oerjan> no wonder ehird wanted to wryte a botte
15:25:39 <ehird> or are you an amnesiac now
15:25:54 <ais523> yes, your upgrade rebooted rutian
15:26:01 <ehird> uh, is that my fault?
15:26:05 <ehird> it was a full upgrade.
15:26:12 <oerjan> ehird: ais523 uploaded himself, you reset him too
15:26:15 <ehird> it did shit to everything. it needed a reboot
15:26:25 <ehird> in fact, iirc the installer didn't even give me a choice
15:27:42 -!- thutubot has joined.
15:27:51 -!- fizziet_ has joined.
15:28:06 <fizziet_> Meh, the previous one disconnectied abruptly.
15:28:22 <thutubot> Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! Hi! ...too much output!
15:29:26 <oerjan> fizziet_ is connecting from a sauna in lahti?
15:29:48 <okleo> actually saunalahti means a lahti containing a sauna
15:29:58 <okleo> lahti is also a word
15:30:02 <fizziet_> Saunalahti ("Sauna bay") is a very common place name in Finland.
15:30:11 <fizziet_> But it's also a finnish ISP/mobile operator.
15:30:18 <ais523> what's with the subsubdomain name? it almost looks like roman numerals but has K in
15:30:36 <MizardX> +ul ((.)S:^:^:^:^:^:^:^:^:^:^):^
15:30:40 <thutubot> ................................................................................................................................ ...too much output!
15:30:40 <fizziet_> I'm not sure; but I think the roman numerals correspond directly to the IP.
15:31:15 <fizziet_> Or "directly" in the sense that it's some base address + the roman numeral number.
15:31:18 <fizziet_> Not sure what they use K for, though.
15:32:19 <ais523> K = 10000 would make sense, possibly
15:33:15 -!- fizziet has quit (Read error: 60 (Operation timed out)).
15:34:25 <oerjan> okleo: you are not fooling anyone all norwegians know Lahti is a place where they do skiing
15:34:42 <oerjan> so clearly you would want a lot of saunas for afterwards
15:35:45 <okleo> well naturally there are saunas in lahti, finnish homes are basically saunas with a second oven for cooking and another set of seats for sleeping.
15:36:06 <okleo> i just meant the composite noun doesn't mean Lahti, that's just not how composites nouns fork.
15:37:01 <oerjan> true that. i've found it weird that greek nouns sometimes _do_ fork that way
15:37:04 * okleo throws some water on the oven, getting cold
15:37:52 <okleo> hmph, god i hate it when things refuse to happen
15:38:41 <oerjan> e.g. philosophia = love of knowledge, from philos = lover and sophia = knowledge or something like that
15:39:46 <okleo> hmm, i've often wondered about that, actually
15:40:29 <oerjan> i haven't found a reference for this though
15:55:02 -!- fizziet_ has quit ("Disconnecting to conserve the phone's battery.").
16:13:42 -!- sebbu has joined.
16:29:34 <okleo> the university is not good for me, i'm actually making an uml diagram for my own purposes atm.
16:29:37 -!- okleo has changed nick to oklo.
17:01:36 -!- Ilari has joined.
17:02:11 -!- kar8nga has joined.
17:12:19 -!- KingOfKarlsruhe has joined.
17:23:34 -!- kar8nga has left (?).
17:25:27 -!- Asztal has joined.
17:46:26 -!- olsner has joined.
17:56:33 <oerjan> Ceci n'est pas un anniversaire.
17:56:50 <ais523> oerjan: what in particular isn't it a birthday of?
17:57:25 <oerjan> http://en.wikipedia.org/wiki/Ren%C3%A9_Magritte
17:57:49 <ais523> oerjan: I'm aware of the whole ceci n'est pas thing
17:57:54 <ais523> is it Magritte's birthday, I wonder?
17:58:22 * oerjan recommends clicking the link
18:02:49 -!- bsmntbombdood has quit (Read error: 110 (Connection timed out)).
18:03:30 -!- bsmntbombdood has joined.
18:20:57 -!- Mony has joined.
18:22:07 <oerjan> Ceci n'est pas un plop
18:23:39 <olsner> norwegians, norwegians
18:23:51 <oerjan> It also isn't Magritte's 110th birthday
18:38:54 -!- sebbu2 has joined.
18:49:02 <AnMaster> this is not a line not sent on a non-existing protocol not called irc
18:49:24 <ais523> I think that's a true statement...
18:49:53 <AnMaster> ais523, so tell me if it is true or not
18:50:18 <ais523> so it wasn't sent on a protocol not called IRC
18:50:24 <ais523> and it wasn't sent on a protocol that didn't exist
18:50:31 <AnMaster> also the master of nested negations would be Marwin in HHGTG
18:50:52 <ais523> it collapses to not ((line sent not on IRC) and (line sent on nonexistent protocol))
18:51:16 <AnMaster> "this is not a line not sent.."
18:51:24 -!- sebbu has quit (Connection timed out).
18:51:24 -!- sebbu2 has changed nick to sebbu.
19:26:39 -!- Judofyr has joined.
19:29:16 -!- Asztal has quit (Read error: 110 (Connection timed out)).
19:49:24 -!- Corun has quit (Remote closed the connection).
19:49:38 -!- Corun has joined.
20:07:40 -!- oerjan has quit ("leaving").
20:08:58 -!- ehird has left (?).
20:08:59 -!- ehird has joined.
20:29:01 -!- ab5tract has joined.
20:45:13 <Deewiant> http://blog.jaoo.dk/2008/11/21/art-and-code-obscure-or-beautiful-code/ -- history of programming languages, 20 minutes in and they've already covered even befunge and piet :-)
21:09:20 -!- Asztal has joined.
21:17:32 <Deewiant> they even present a shakespeare program as a play :-))
21:18:41 <ais523> someone should convert that video to a format I can actually see
21:25:43 <AnMaster> ais523, this url may work http://e6.video.blip.tv/0720006203922/Krab-50In50ByRichardPGabrielAndGuyLSteele278.avi
21:26:18 <ais523> actually, I just realised that an hour-long avi is probably not good to download
21:26:28 <ais523> given that IIRC it's uncompressed
21:28:43 <AnMaster> ais523, there are other formats there
21:29:00 <AnMaster> ais523, http://blip.tv/file/1472720/
21:29:54 <Asztal> AVI should be compressed
21:30:11 <ais523> none of the formats there are ones I can easily view, really
21:30:27 <Deewiant> What /can/ you easily view? O_o
21:30:42 <ais523> probably that's about it, tbg
21:30:50 * Deewiant points to http://www.videolan.org/
21:31:04 <ais523> but they changed the extension to avoid ambiguity with Ogg Vorbis
21:31:04 <AnMaster> ais523, what is ogg vorbis then?
21:31:14 <ais523> .ogg for Vorbis, .ogv for Theora
21:31:42 <AnMaster> still ogv/otg would have made more sense
21:36:12 <AnMaster> Deewiant, I prefer flac for my music
21:36:29 <AnMaster> much better sound quality really
21:36:39 <AnMaster> and about half the size of the raw format
21:37:55 <Deewiant> For purity's sake I keep stuff in a lossless format when possible even though I can't tell the difference between it and lossy
21:46:29 -!- Asztal has quit (Read error: 104 (Connection reset by peer)).
21:46:48 <oklo> have to take a break to make me some noodles
21:47:07 <oklo> also i wish i wasn't so goddamn tired, i miss some of the awesome jokes :P
22:17:51 -!- KingOfKarlsruhe has quit (Remote closed the connection).
22:25:40 -!- Sgeo has joined.
22:56:53 -!- Asztal has joined.
23:02:54 -!- nooga has joined.
23:14:34 <fizzie> That's a perfectly valid greeting.
23:15:58 <fizzie> No, "moronnääs". My only connotation for "nt" is "no text".
23:16:10 <fizzie> Or Windows NT, maybe. New Technology, was it?
23:16:16 <oklo> i usually greet people with a random syllable if i know them well enough to stop for a chat
23:16:31 <oklo> for people i just greet and walk by, i usually just open my eyes real wide
23:16:33 <fizzie> I usually greet people with "arrr".
23:16:35 <oklo> and stare at them
23:16:54 <oklo> ...some don't understand that's a greeting, i've been told i'm scary
23:17:22 <MizardX> http://paste.eso-std.org/ is broken :(
23:17:37 <oklo> pb.vjn.fi for all your pasting needs
23:18:46 <fizzie> There is quite a lot of those things around.
23:18:53 <MizardX> Was thinking of uploading a new version of my redivider parser. This one isn't limited by the python stack.
23:19:02 <oklo> well it takes 5 minutes to code and everyone has a place to put one int
23:19:07 <oklo> and everyone needs them
23:19:14 <oklo> so i wouldn't say that's surprising
23:19:18 <AnMaster> fizzie, yes I plan to create a list
23:19:23 <fizzie> I just paste things to .txt files in my ~/www/ ... maybe I should join the new millennnnnium some day.
23:19:29 <AnMaster> the definite guide to five star pastebins
23:19:53 <oklo> vjn.fi/pb is the best i've seen
23:20:05 <oklo> can't imagine a better one really
23:20:14 <ais523> http://vjn.fi/pb <--- ais523 gives themself a link to click on
23:20:25 <nooga> i do not have left eye
23:20:42 <AnMaster> nooga, is that random or a true statement?
23:21:00 <oklo> syntax highlighting isn't that useful
23:21:08 <AnMaster> for competent syntax highlight of erlang
23:22:04 <oklo> heh i read "enjoy your paste"
23:22:11 <oklo> hmm. not sure that would've been weird
23:22:21 <oklo> probably would
23:22:43 <nooga> wind + broken glass killed my eye!
23:23:06 <oklo> are you serious?=
23:23:21 <oklo> i thought that was a pirate joke
23:23:37 <nooga> i look like a pirate now
23:23:47 <oklo> is it like, officially broken?
23:23:49 <nooga> but probably they will manage to repair that
23:24:03 <nooga> but it fscking hurts
23:24:12 <oklo> i've always said the wind is of the devil.
23:24:56 <nooga> i was just walking down the street during this ff.... hurricane
23:25:00 <oklo> that lisp pastebin doesn't give you the link, huh?
23:25:33 <oklo> oh, you have to give a nickname
23:26:10 <oklo> that's the problem with pastebins, i just want to share stuff as quickly as possible, almost every pastebin has their own useless additions.
23:26:15 <oklo> that's just stupid
23:27:12 <oklo> i was contemplating buying me a pair of those big glasses snowboarders use
23:27:27 <oklo> because my eyes hurt for hours after being in the wind for a wihle
23:27:44 <oklo> and third random topic, fizzie, how went the talk?
23:28:19 <oklo> did you get to strip naked?
23:28:33 <oklo> did you start with "moronnääs"?
23:28:44 -!- Asztal has quit (Connection timed out).
23:28:51 <nooga> i can't aim right keys on the keyboard
23:28:57 <oklo> i should try doing something. i've been doing nothing all day
23:30:22 <oklo> there was some experiment where this guy used some kinda oculifiers to turn the world into a mirror image, then lived through that for a month and learned to operate perfectly with it
23:30:25 <oklo> i want to try that
23:30:51 <fizzie> oklo: No stripping, no "moronnääs" (although the location would've been appropriate for that, I guess); actually there was nothing especially interesting about how it went. Not too much people in the audience, pretty much only the other presenters and a couple of random colleagues of theirs.
23:31:21 <oklo> fizzie: did you get moneys?
23:31:35 <oklo> olsner: it's a greeting used in a town in finland
23:32:04 <olsner> this town has its own greeting?
23:32:18 <oklo> olsner: all towns have their own words
23:32:35 <fizzie> oklo: No. Well, I did get to count it as a work-day. I don't think people usually tend to pay for seminar-presentation-style talks; real lecture-style things are a different matter, of course.
23:32:48 <oklo> nooga: i suggest learning sentences you can write with your keyboard :P
23:33:39 <oklo> fizzie: what is it you do for work if you don't mind my curiousity?
23:33:42 <oklo> nooga: it's hyvää päivää
23:34:00 <oklo> my ä's look weird, nnscript just won't learn.
23:34:33 <fizzie> oklo: Speech recognition research, pretty much.
23:34:35 <oklo> nooga: in hyvaa paivaa, all the a's have umlauts
23:34:59 <nooga> but i was too lazy to open char map
23:35:02 <olsner> without the umlauts it's just gibberish :)
23:35:48 <oklo> surprisingly easy to read considering finland very rarely has any link between a's and ä's
23:35:52 <nooga> in Polish internet speak almost nobody uses ąóńćęłżźś
23:36:18 <fizzie> oklo: The slides for today's (well, yesterday already) thing are at http://zem.fi/~fis/tampere08.slides.pdf but I guess they don't make much sense without the presentation part.
23:36:31 <nooga> sometimes it becomes hilarious
23:37:03 <oklo> so far so good
23:37:08 <oklo> The \standard" mel-frequency cepstral coecient (MFCC)
23:37:09 <oklo> based feature extraction is relatively sensitive to noise
23:37:13 <fizzie> 'ä' and 'a', or 'ö' and 'o' are pretty different sounds.
23:37:21 <oklo> HUH? please repeat that i must have misheard
23:37:29 <olsner> polish is weird... way too many 'sz'-sounds
23:37:35 <fizzie> Heh, seems like there's a "ff" ligature there.
23:37:55 <oklo> fizzie: i doubt that'd make much sense to me with the presentation either ;)
23:38:08 <olsner> 'ffi' unless coefficient is misspelled in the original text :)
23:38:32 <oklo> dunno if you actually require me to know something about the standard mel-frequency cepstral coefficient
23:38:46 <fizzie> Everyone there was speech people, I didn't much explain the basics.
23:38:53 <nooga> "zrób mi łaskę" means "do me a favour", but "zrob mi laske" (which should mean the same) means "do me a blowjob"
23:38:55 <fizzie> MFCCs are explained later on, though.
23:39:44 <fizzie> For some values of "explained", anyway.
23:40:02 <olsner> nooga: heh, maybe you should asciiize your orthography :)
23:40:09 <fizzie> It doesn't mention what's the point, but...
23:40:18 <oklo> window the input signal into overlapping frames
23:40:23 <oklo> so, you have some string as input
23:40:28 <oklo> and you kinda cut it into pieces
23:40:30 <oklo> but they overlap
23:40:41 <fizzie> Well, it's audio data what we get as input.
23:40:51 <oklo> i was just generalizing
23:41:19 <olsner> like, wroclav -> vrotswaf (obviously the choices are arbitrary though)
23:41:21 <fizzie> Yes, we cut something like 16 millisecond frames out of it, and for each frame advance the starting point by 8 ms, so they end up overlapping.
23:41:34 <oklo> estimate the log of the amplitude spectrum... what's the definition of logarithm for a spectrum..?
23:42:08 <fizzie> It's just the logarithmic value for the amplitudes.
23:42:21 <nooga> olsner: wrocław, it is
23:43:04 <fizzie> The word "cepstrum" is a bit funny, though.
23:43:05 <olsner> yeah, it doesn't end with a 'v' ... don't know where I got that from
23:43:11 <oklo> what does it mean
23:43:19 <fizzie> It's sort-of the inverse of "spectrum".
23:43:30 <fizzie> So the "spec" part is reversed to "ceps".
23:43:33 <oklo> yeah okay that would've been one helluva typo
23:43:39 <nooga> i woult write it something like vrotslaf
23:43:43 <fizzie> And filters in the cepstral domains are called lifters, for the same reason. :p
23:43:48 <oklo> maybe possible if writing the text as a befunge program outputting it
23:43:58 <olsner> a 'cepstrum' is a spectrum transformed back into the time-domain?
23:44:45 <nooga> i should go to sleep
23:44:49 <fizzie> olsner: Well, it's the log-amplitude spectrum transformed again; it's not exactly "time-domain" any more at that point.
23:45:17 <fizzie> Oh, and the "X axis" in the cepstrum is the quefrency. :p
23:45:30 <oklo> whose terminology is this?
23:45:35 -!- nooga has quit ("Lost terminal").
23:45:47 <fizzie> Just see http://en.wikipedia.org/wiki/Cepstrum for example.
23:45:57 <fizzie> There's liftering and quefrency mentioned.
23:46:09 <fizzie> I'm not sure if there are other punny names.
23:46:53 <fizzie> I haven't seen "alanysis" anywhere, though. Well, except that wikipedia article.
23:47:03 <olsner> "the power cepstrum (of a signal) is the square of the Fourier transform of the logarithm of the squared magnitude of the Fourier transform of a signal"
23:47:54 <fizzie> In MFCC's the log-spectrum is first warped to the Mel scale, which is a sort-of a perceptually motivated logarithmic thing.
23:47:54 <oklo> zifzie is an epsech ocergnition eserarcher
23:48:25 <fizzie> I think the original Mel scale was derived by playing sine tunes to people and having them to classify which are "equally distant".
23:48:56 <fizzie> And then approximating that a bit.
23:49:11 <oklo> people who have trouble hearing notes right?
23:50:22 <oklo> it's just a logarithmic scale +- failure to hear notes right, in case you know anything about music
23:50:41 <oklo> well i guess it's not about how much you know
23:51:16 <fizzie> It's what "sounds" equally distant; and it's supposedly not logarithmic through the whole frequency range.
23:51:55 <oklo> hmm. i would just tell them the interval, not sure what equally distant could mean apart from that :P
23:52:00 <olsner> I'm getting curious... maybe I should rebuild my haskell spectrum analyser into a cepstrum alanyzer
23:53:11 <oklo> well, guess if it were log scale +- small error, it would not be mentioned there
23:53:37 <fizzie> It's just that from a given a frequency X, the "distance" between (X/2, X) and (X, 2*X) doesn't necessary sound like it's the same.
23:54:01 <fizzie> Or so they tell me, anyway. Haven't done any listening tests myself.
23:54:23 <oklo> if i hear X/2 and X, i hear an octave. i'm assuming most people do, even though they may not have a name for the interval
23:55:14 <oklo> if i hear X and 2X after that, i hear another octave, unless the notes get too high to hear right, but i doubt that's where it's supposed to start deviating from log scale
23:56:39 <oklo> but maybe the western note system isn't a very natural one, and people who aren't fluent in it have the more natural system
23:57:28 <oklo> of course i guess i could just wp it
23:57:41 <fizzie> The wikipedia article is not terribly informative.
23:58:07 <fizzie> Anyway, the approximation of Mel scale used by everyone is completely logarithmic, so...
23:58:51 <oklo> hmm. i want some coke
23:59:10 <oklo> ...wow i just realized something
23:59:10 <fizzie> The original formulation (if I recall correctly) had a linear part for low (<500 Hz) frequencies, and then logarithmic above that; supposedly it's related to how the ear does it's magic.
23:59:37 <oklo> if i went out, and bought coke, NO ONE COULD STOP ME :D
23:59:52 <oklo> i can just do it
23:59:55 <oklo> life is awesome