00:50:45 -!- lindi- has quit (niven.freenode.net irc.freenode.net).
00:52:04 -!- lindi- has joined.
00:56:56 -!- kipple_ has quit (Read error: 110 (Connection timed out)).
00:58:12 <Razor-X> I need to learn TI-ASM and code BF for that.
00:59:01 <calamari> that must be for the newer TI's
00:59:19 <calamari> I wrote my BF interp on the ti-82
00:59:25 <pikhq> I need to learn RPL.
00:59:44 <pikhq> Reverse Polish Lisp.
01:00:04 <calamari> I need to learn ARM assembly language
01:00:05 <pikhq> It's the builtin language on my calculator.
01:00:25 <pikhq> Although I might want to bother figuring out how to get GCC-HP working, isntead.
01:00:40 <Razor-X> That's awesome. I want Lisp on my calculator ;-;
01:00:47 <Razor-X> Well.... it's a pain on that horrible keyboard.
01:00:59 <pikhq> Also, the HP ROM is free sofware. . .
01:01:08 <Razor-X> We're required to have TIs.
01:01:13 <calamari> what's wrong with the ht keyboard?
01:01:26 <Razor-X> I'm not sure about HP's keyboard, but the TI keyboards are uggh.
01:01:40 <Razor-X> I should grab that one implementation of Scheme for the Palm.
01:03:09 -!- ihope_ has quit ("http://tunes.org/~nef/logs/esoteric/06.08.09").
01:08:11 <Razor-X> Does anyone know where one can get old laptops on t3h cheap?
01:14:22 <RodgerTheGreat> if you live anywhere near a university, try to find the location of their "Property Disposition Center".
01:14:52 <Razor-X> I live sorta near Stanford.
01:15:07 <RodgerTheGreat> I went to U of M's dispo center for several years- always a good place to look for cheap, obsolete parts.
01:15:35 <Razor-X> I need to slap a form of GNU Linux on the thing. I need to be in perfect control of the environment for the project, and for that I need GNU Linux.
01:15:44 <RodgerTheGreat> I bought a 200mhz powerbook for $75 there once- pretty good deal considering the battery was in perfect condition.
01:15:51 <Razor-X> Hopefully it can support a 2.6 kernel so I can play with ALSA's OSS emulator layer.
01:16:23 -!- GregorR-L has joined.
01:16:46 <Razor-X> Well, I'll probably be needing an X server of some sort (XVesa is my guess), GNUPlot to make life easier, and if possible, the capability to run Chicken.
01:17:06 <RodgerTheGreat> you sure Bochs or the like wouldn't suit your purposes better? Emulating a machine probably wouldn't be much slower than using an obsolete laptop, and you'd have more configuration control.
01:17:42 <Razor-X> Nope. I just need a show for the science fair people.
01:18:07 <Razor-X> Does QEmu work in Windows?
01:18:20 <pikhq> Razor-X: One can do ALSA on 2.4.
01:18:40 <calamari> Razor-X: yeah Qemu works in windows
01:18:45 <Razor-X> pikhq: But I need the OSS Wrapper.
01:18:58 <pikhq> Razor-X: And the OSS wrapper also works in 2.4.
01:19:28 <pikhq> ALSA, after all, was originally just some additional 2.4 modules (xor 2.4 patch). ;)
01:20:01 <Razor-X> I hope mplayer can work too, so I don't have to write C wrapper code.
01:20:41 <calamari> Razor-X: http://www.h7.dion.ne.jp/~qemu-win/
01:21:06 <Razor-X> I'm going to be writing everything.
01:21:36 <Razor-X> Although I'll need to graduate from the transfer function ;).
01:22:16 <Razor-X> Heh. I had to argue my dad into submission to believe it though.
01:22:28 <Razor-X> Finally he says ``You're too bullheaded to refuse, so whatever.''
01:24:04 <RodgerTheGreat> neural nets are really one of the best examples of pure success in AI research.
01:24:22 <RodgerTheGreat> by designing your net and training algos properly, they can achieve amazing results.
01:24:22 <Razor-X> The thing is, I noticed while reading the DSP books that convolution is just the application of a weighted average on the input signal. I was thinking of having the NN's weights emulate the convolution signal (well, that'll be one of the experiments, I'll be abstracting it to many different levels) and see if it can beat the algorithms already written.
01:25:24 <RodgerTheGreat> advancing the state of the art is a surefire way to win in a contest.
01:25:43 <RodgerTheGreat> speaking of which, have you received any submissions yet?
01:26:03 <Razor-X> I've been *very* busy this weekend up till now, and I'll be busy in a few minutes for test studying.
01:26:18 <Razor-X> pikhq submitted his a long time ago, other than that, I haven't had the time to check.
01:26:45 <Razor-X> I have AP Physics formulas to memorize which is easy, but a whole bunch of US historic people to remember which is annoying and useless.
01:27:00 <RodgerTheGreat> I'm heading a team in my dorm that's entering MTU's "cardboard boat" competition. It'll probably be soaking up *all* my free time for the next week.
01:27:05 <Razor-X> I'll be abstracting on a whole bunch of levels. From the ``magic'' level (I give you blah-very-noisy-signal, I want blah-un-noisy-signal) to the level of emulating the convolution signal using the weights.
01:29:02 <Razor-X> It'll be a *HECK* of a lot of work. I'll be writing GNUPlot wrappers, NN base classes, NN topologies, signal classes and convenient signal manipulation functions, and this is before any of the actual testing of the experiments.
01:31:18 <Razor-X> That's why I wanted mplayer. I don't want to have to write wrapper code to set the parameters for playing with audio raw.
01:31:18 <RodgerTheGreat> hm. I haven't coded a proper neural net system since my BASIC days. perhaps I should build a new one with Java...
01:31:39 <Razor-X> Lisp has a particular advantage in that functions are datatypes. I'll be exploiting this for NN writing.
01:32:53 <RodgerTheGreat> replacing current chip designs with neural nets sounds like it'll be a trick, but the technique could improve sound emulation tech... *running* a neural net doesn't take much CPU power once it's properly evolved and tweaked.
01:33:24 <Razor-X> Well, if the weigths can produce convolution signals, then you don't even need an NN.
01:33:35 <Razor-X> (This is one level of abstraction, of course.)
01:33:55 <Razor-X> Have an NN produce the weights beforehand, then make a signal using that and store it in the particular filters you want to implement.
01:34:06 <RodgerTheGreat> yes, then you could turn the single-layer net into a matrix operation.
01:35:15 <RodgerTheGreat> heh. you might actually be able to use such a technique to have a GPU emulate a soundcard- GPU's are built primarily around vector processors- dedicated matrix multipliers.
01:35:18 <Razor-X> The best thing about NNs is that, it lets high schoolers play algorithm with the big mathematicians.
01:35:43 <Razor-X> I'm sticking to sound simply for the purposes of the project. If I have the time to do graphics research too, I'll add that in later.
01:36:23 <Razor-X> But raw audio is A) Really easy to get to play in Linux B) Based on very simple math C) Easy to sample.
01:36:50 <RodgerTheGreat> well, I'm just saying the GPU could be used to *accelerate* neural net evaluation, if you needed them for a serious purpose. I'm not sure how feasible simulating GPU's with NN's would be...
01:38:17 <RodgerTheGreat> the other advantage of audio is that it depends less on precision than most types of calculations a computer could be used to do. Audio can be processed in a "good enough" fashion with little harm.
01:40:36 <Razor-X> I'll also have to be creative in my sampling... the sample may be *particularly* bad, for illustrative purposes ;).
01:42:07 <Razor-X> I've heard that raw voice samples are pretty awful.
01:42:49 <RodgerTheGreat> trying to build a NN-based vocoder could be interesting. input nodes would take the raw audio signal, and output nodes could extract formants... then you could build a vox around the opposite principle, using outputs to modulate a pure sine sample or somesuch thing...
01:43:29 <Razor-X> My dad couldn't understand why you have multiple inputs to the NN when it's only one signal :D
01:43:33 <RodgerTheGreat> I need to figure out how to do raw audio I/O from java.
01:43:49 <pikhq> Link to a C++ library.
01:44:18 <RodgerTheGreat> might work. A more platform-independent solution would be preferable, but hey..
01:44:24 <Razor-X> RodgerTheGreat: Tritonius.
01:45:15 -!- puzzlet has quit (Read error: 145 (Connection timed out)).
01:45:21 <Razor-X> I've been looking into Java for the AP test, so jah.
01:47:03 <RodgerTheGreat> although, it looks like sun's API replicates most of the functionality.
01:48:30 <Razor-X> I just need to set sample rate, and sample format.
01:52:11 <RodgerTheGreat> are you going to be comparing your nets to existing algorithms, performance-wise, or is the experiment mainly to prove that NN's are a viable solution?
01:53:08 -!- puzzlet has joined.
01:53:21 -!- puzzlet has quit (Client Quit).
01:53:34 -!- puzzlet has joined.
01:53:44 <RodgerTheGreat> you should try pounding some XM's or MOD's through your filters- it'd be interesting to see if the clearer synthetic patterns worked better, or impaired the nets.
01:56:42 <RodgerTheGreat> anyway, if there's anything I can help out with, just let me know.
01:57:06 <Razor-X> But aren't XMs and MODs already cleaned up?
01:58:11 <Razor-X> I need something with noise and make it un-noisy.
01:58:42 <RodgerTheGreat> well, they're synthesized in the first place. Comparing a digitally generated one with a loopback-recorded one could get you some clear metrics for cleanup.
01:59:04 <Razor-X> In fact, that's a very good idea.
01:59:33 <Razor-X> I'll use the loss playing the MOD or XM, resample that, and compare that to the ideal signal.
01:59:38 <Razor-X> Or any such audio sample, actually.
02:00:15 <RodgerTheGreat> the trick is to use some type of sample that you can generate a "perfect" control sample for, such as MIDI, SID, MOD, XM, etc.
02:00:34 <Razor-X> Well, I can tell mplayer to generate a perfect sample for me :3
02:02:19 <RodgerTheGreat> man, I have this song stuck in my head: http://rodger.nonlogic.org/music/Rebb_-_Devious.mod
02:03:53 <Razor-X> Oh, and thank calamari for the inpsiration for a lot of this.
02:04:26 <Razor-X> And thank you for the NN inspiration :D.
02:07:00 -!- oerjanj has joined.
02:28:43 <oerjanj> just reading Lambda the Ultimate
02:33:50 <oerjanj> i see a real Old Hat has joined the wiki...
02:37:35 -!- GregorR-L has quit (Remote closed the connection).
02:39:06 <RodgerTheGreat> referring to the "Prehistory of Esolangs" page? I thought it was pretty interesting.
02:45:51 <RodgerTheGreat> http://esoteric.voxelperfect.net/wiki/Prehistory_of_Esoteric_Languages
02:47:55 <RodgerTheGreat> I'm running off of a T3 atm, so any delay whatsoever feels weird.
02:48:32 <RodgerTheGreat> A couple weeks ago, my roommate managed a 14megs/sec download rate. ^_^
02:48:47 <Razor-X> 1:27 and the page isn't loading.
02:51:13 <oerjanj> hi razor-x, did jix ever get hold of you? from reading the logs you seemed to behave like Superman and Clark Kent...
02:53:04 <Razor-X> On weekdays, I'm away for a long time.
02:53:09 <Razor-X> This weekend I've been very busy.
02:54:14 -!- CXI has quit (SendQ exceeded).
02:55:06 * pikhq should give Dimensifuck a ooD memory as well. . .
02:57:14 <oerjanj> i think that was suggested already, possibly by me
02:59:31 <pikhq> I'm saying I should actually do so.
03:53:35 <Razor-X> You have been shamed GregorR.
03:54:04 <Razor-X> I met someone who is in almost as many channels as you're in, squared.
03:59:58 <lament> 76 is not _quite_ almost 144.
04:01:02 <Razor-X> In respect to GregorR it is.
04:08:44 <Razor-X> Depends on his advertised value.
04:36:57 -!- GregorR-L has joined.
04:41:57 -!- CakeProphet has quit (Read error: 145 (Connection timed out)).
04:46:37 <oerjanj> damn, i'm trying that puzzle and when googling to find information on the hint on page thirteen, i end up at a cheat sheet for the puzzle
04:47:20 <oerjanj> just as well i would never have figured out that hint otherwise.
05:12:34 -!- Razor-X has quit (Read error: 104 (Connection reset by peer)).
05:14:38 -!- Razor-X has joined.
05:32:36 -!- oerjanj has quit ("Leaving").
05:33:27 <wooby_> anyone looked into the shortest possible BF quine
05:34:57 <GregorR> Razor-X: Y'know, I'm not in a bunch of channels for novelty value.
05:35:10 <GregorR> Razor-X: And is this person in so many channels, oh, say, ChanServ? ;)
05:35:25 <Razor-X> All of you, whois a person called ChrisWhite.
05:39:50 <GregorR-L> I suppose if you join #gentoo* and #suse* you can increase your channel count exponentially, but i don't consider that channel list particularly legit.
05:40:51 <wooby_> too bad there isn't a #gentoo-ppc-embedded-laptop-Anime
05:49:07 -!- anonfunc has joined.
05:56:31 -!- Sgeo has quit (Connection timed out).
06:11:29 -!- calamari has quit ("Leaving").
06:51:49 -!- CXI has joined.
07:10:43 -!- Arrogant has joined.
07:40:20 -!- GregorR-L has quit ("Leaving").
07:43:43 -!- CXI has quit (Read error: 104 (Connection reset by peer)).
07:43:58 -!- CXI has joined.
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:01:24 -!- GregorR-W has quit (Remote closed the connection).
08:02:16 -!- GregorR-W has joined.
08:59:30 -!- anonfunc has quit.
09:05:24 <Razor-X> CLOS is pretty interesting.... but it seems useless when you aren't doing heavy OOP.
09:08:01 <Arrogant> Which is almost all the time in Lisp
09:08:15 <Razor-X> I'll be using a structure then.
09:08:30 -!- nooga has joined.
09:09:50 <Razor-X> Ugggh. I don't want to write an entire class system myself -_-''.
09:12:54 <Razor-X> Maybe I will be using Tiny CLOS :D.
09:21:40 <Razor-X> Tiny is the Scheme version.
09:25:13 <Razor-X> Creating code for NN classes.
09:28:01 <Razor-X> For science project stuff.
09:29:08 <nooga> i jutst woke up :>
10:07:58 -!- Sgeo has joined.
10:08:48 -!- Sgeo has quit (Client Quit).
10:29:00 -!- Arrogant has quit ("Leaving").
11:10:19 -!- lindi- has quit (niven.freenode.net irc.freenode.net).
11:10:46 -!- lindi- has joined.
12:48:35 -!- kadmut has joined.
14:43:58 -!- calamari has joined.
15:02:27 -!- tgwizard has joined.
15:16:05 -!- jix has joined.
15:35:25 <Razor-X> jix: I'm going to school now, please PM me the business you have.
15:35:51 <Razor-X> You wanted to talk to me, right?
15:36:05 <jix> not anymore.. it was about the contest but it is over now right?
15:36:16 <Razor-X> I've been real busy. Sorry.
16:15:06 -!- nooga has left (?).
16:20:59 <jix> RodgerTheGreat: may i enter the contest after the deadline with the langauge linguine...
16:21:13 <jix> hmm until midnight? which time?
16:21:21 <jix> could be possible to do them
16:23:26 <RodgerTheGreat> Razor-X didn't like the idea of people registering after the fact. Her main idea was that some languages could give an unfair advantage, with the goals in mind.
16:24:30 <jix> hmm which langauge of the allowed one could allow me to complete the challenges until midnight
16:24:41 <RodgerTheGreat> I'm going to give you two options: 1) try to complete something in time with a language someone has already registered with, or 2) go with linguine and see what you can pull off, knowing that you probably won't end up being eligible for more than an honorable mention
16:25:46 <RodgerTheGreat> well, we have BF, Unlambda, Befunge, TheSquare, INTERCAL, and malbolge.
16:26:04 -!- kipple_ has joined.
16:26:33 <RodgerTheGreat> I don't think she'll be too happy with me for letting you do this, but I'm not sure how many entries we'll be having in the first place.
16:29:42 <jix> wow! we have a female esoteric programmer!
16:32:26 <jix> no one wants to do malbolge?
16:32:48 <jix> i don't like unlambda :(
16:33:06 <jix> (compared to lazy-k)
16:33:22 <jix> but on the other side... lazy-k is kinda boring compared to unlambda
16:33:25 <RodgerTheGreat> hey, if you can solve one of those problems in malbolge before midnight, I can tell you for a fact that you'll win the competition hands-down.
16:34:05 <kadmut> some guys did metaprogramming in malbolge using genetic algorithms
16:34:31 <RodgerTheGreat> the cryptanalysis was really the breakthrough development, though.
16:56:02 <RodgerTheGreat> so, jix- do you think you're going to try building a submission?
16:56:17 <jix> hmm not sure
16:56:25 <jix> not in one of the languages mentioned
16:56:51 <RodgerTheGreat> I wouldn't blame you for wanting to back out- you're really down to the wire on time one way or another.
16:57:16 <RodgerTheGreat> we can try holding a followup competition in a few months if there's interest.
16:58:21 <jix> i'll try to get a linguine solution
18:12:01 -!- GregorR-W has changed nick to \x1B[2J.
18:12:49 -!- \x1B[2J has changed nick to GregorR-W.
18:29:20 <jix> i'm starting right now
18:38:00 <jix> i wont join
18:38:05 <jix> i'll wait for the next competition
18:39:42 * jix codes game instead of esolangs
18:40:28 <jix> current creenshot: http://img158.imageshack.us/img158/6231/bild10gk1.png
18:41:53 <GregorR-W> 2D? 2D is soooooooooooooooooo 2003.
18:42:13 <jix> well this is more like 1980
18:42:14 <GregorR-W> All the cool kids use 3D even for things where 3D makes no sense.
18:42:41 <jix> indeed graphic wise i used two 1980 shooters as a uhm ..... dunno the word right now
18:42:43 <GregorR-W> How can you use Mac OS X and not use OpenGL for 2D graphics? ;)
18:42:56 <jix> GregorR-W: i'm using SDL
18:43:05 <jix> because i want to port it to some things
18:43:10 <jix> and SDL makes that pretty easy
18:43:57 <jix> GregorR-W: i don't believe you
18:44:23 <jix> GregorR-W: what do you think about the gfx?
18:44:32 <GregorR-W> I like the interlacing effect, very retro.
18:44:48 <jix> yeah but thats only on 640x480...
18:45:08 <jix> the game is/gets written for 320x240 but it's too small on a 19" so i added that mode for 640x480
18:45:33 <GregorR-W> How 'bout if you try to expand it to, say, 1024x768?
18:45:42 <jix> why should i?
18:45:48 <jix> would be possible
18:45:57 <GregorR-W> So that I can play it full-screen X-P
18:46:10 <jix> but atm it's hardcoded to either 640x480 or 320x240 with preprocessor macros
18:46:18 <jix> i'll take care bout that later
18:46:32 <jix> atm it's possible to move the screen out of the window and everything
18:46:38 <GregorR-W> (Also, software should be able to scale to screen size, otherwise they're forever stuck in the present, and will be worthless in a few years when 640x480 = postage stamp :) )
18:48:41 <pikhq> Just do your graphics in SVG.
18:49:04 <jix> pikhq: no way!
18:49:07 <pikhq> Maybe have it prerender the SVG upon installation.
18:49:16 <jix> pikhq: that won't look nice
18:49:30 <jix> pikhq: because this is a 1980 style game
18:49:46 <jix> and 1980 style games had every pixel set by hand... and that gave them a special look and feel
18:49:50 <jix> and i'll do the same
18:50:26 <jix> i WANT this blocky icancountthepixels feeling....
18:51:05 <jix> pikhq: yeah
18:51:36 <jix> GregorR-W: linear? nearest neighbor + scanlines
18:51:51 <GregorR-W> jix: Doesn't that make it appear less pixely?
18:52:03 <jix> GregorR-W: no.... linear make it a bit smooth
18:52:28 <GregorR-W> Oh ... I guess I'm misremembering :P
18:52:28 <jix> nearest neighbor is just paint every pixel n times....
18:52:37 <jix> linear is.. well linear interpolation
18:52:48 <GregorR-W> Not sure why "linear" popped into my head.
18:52:53 <pikhq> One can have individual blocks in SVG. XD
18:52:54 <jix> then everything is ok and the world wont explode
18:53:07 <jix> pikhq: O RLY?
18:53:12 <pikhq> <block dim="foo" color="bar" />
19:09:53 <SimonRC> I also read the history of esolangs article and part of the article on TECO
19:11:28 <SimonRC> Brainfuck may be hard to program in, Malbolge may only be usable through genetic algorithms, but if you *really* want your mind boggled, try using Arrowised Functional Reactive Programming: http://www.haskell.org/yampa/AFPLectureNotes.pdf
19:12:12 <SimonRC> Just try to understand the signiture of pswitch: http://www.haskell.org/yale/papers/haskellworkshop02/index.html
19:26:22 -!- fr34k has joined.
19:28:49 <SimonRC> Arrowised Functional Reactive programming
19:29:43 <SimonRC> rpm = rotations per minute
19:30:05 <fr34k> so why does your head spin?
19:30:32 <SimonRC> because I am trying to understand arrowised functional reactive programming
19:31:06 <fr34k> i dont even want to understand that :P
19:31:46 <fr34k> im trying to do a research about pulsejet engines, so i can build one myself
19:42:16 <fr34k> hey is pikhq here?
19:46:31 <SimonRC> pikhq spoke about 50 mins ago
19:54:16 -!- fr34k has changed nick to boblol.
19:56:30 <boblol> i still havnt figured 196 out :P
20:25:53 -!- jix has quit ("Bitte waehlen Sie eine Beerdigungnachricht").
20:26:34 -!- Sgeo has joined.
20:39:46 -!- jix has joined.
20:56:51 -!- CakeProphet has joined.
21:08:35 -!- boblol has quit.
21:36:21 -!- jix has quit ("Bitte waehlen Sie eine Beerdigungnachricht").
23:01:29 -!- anonfunc has joined.
23:08:26 -!- GregorR has quit (Read error: 110 (Connection timed out)).
23:09:54 -!- RodgerTheGreat has changed nick to RodgerTheAfk.
23:18:49 -!- EgoBot has quit (Read error: 110 (Connection timed out)).