00:00:06 <AnMaster> hm, it would be nice with a TM that could rewrite it's own rules
00:00:49 <CakeProphet> so does clisp-for-BEAM completely integrate with Erlang? As in, I can use a clisp module as if it were a normal erlang module?
00:04:14 <AnMaster> CakeProphet, very clearly I said it wasn't clisp but a custom lisp
00:04:41 <AnMaster> but only because it is special crafted for erlang
00:09:57 <CakeProphet> http://wiki.reia-lang.org/wiki/Reia_Programming_Language
00:10:24 <CakeProphet> a Ruby-like language with Python whitespace-significance for BEAM
00:11:28 <cheater99> i never understood what's so special about ruby
00:39:02 <oklopol> a bf program is pretty much directly a turing machine, but a turing machine is not necessarily directly a brainfuck program
00:39:35 <oklopol> the control structure is different, the one of tm's is more flexible
00:40:42 <oklopol> just have a state for each command in the program, and it's pretty obvious what you do in each state, and what state you go to
00:41:43 <oklopol> in case CakeProphet still doesn't know, turing machines have a finite amount of states, and at each time step they may, depending on state and tape content at tape head, change cell content at tape head, move on tape, and change state
00:42:12 <oklopol> in a bf program your states must come in some sort of linear progression with loops, not in arbitrary jumps
00:42:43 <oklopol> this is like while() vs. goto, but in a restricted computation model
00:42:47 <oklopol> well i should go to sleep now
00:49:56 -!- augur has quit (Ping timeout: 265 seconds).
01:12:10 -!- alise has quit (Quit: Leaving).
01:12:15 -!- alise has joined.
01:17:16 <Gregor> http://vetusware.com/download/AT_T%20UNIX%20System%20V%20Release%204%202.1.4/?id=5727 <-- this 1) works and 2) seems to be working on Bochs 8-D
01:17:51 <Gregor> (The boot disk didn't boot in Qemu, but I can always move it over later)
01:24:20 <alise> System V? Who cares?
01:24:48 <alise> It had already succumbed to the crawl of coagulated growth by then. :)
01:25:18 <alise> Gregor: Is pikhq dead for you too?
01:25:27 <alise> Guess I'll just have to compile my own uclibc.
01:25:45 <alise> Flinix must happen.
01:25:48 <Gregor> I have ancient copies of Xenix too, but I haven't made them work in any emulator, not even MESS.
01:26:05 <alise> Now Gregor must ask me what Flinix is.
01:26:21 <alise> Now Gregor must ask me "What is Flinix?".
01:26:44 <alise> Now Gregor must ask me What is Flinix?.
01:28:58 -!- oerjan has quit (Quit: Good night, and good luck).
01:30:11 -!- GreaseMonkey has quit (Quit: I'm using NO SCRIPT WHATSOEVER - Download it at file:///dev/null).
01:33:17 -!- Gracenotes has quit (Ping timeout: 240 seconds).
01:38:07 -!- Oranjer has joined.
01:41:56 -!- Mathnerd314 has joined.
01:44:09 <CakeProphet> The objectives of software visualizations are to support the understanding of software systems (i.e., its structure) and algorithms (e.g., by animating the behavior of sorting algorithms) as well as the analysis of software systems and their anomalies (e.g., by showing classes with high coupling).
01:44:19 <CakeProphet> I wonder why software companies are OBSESSED with sorting algorithms.
01:45:47 <alise> http://www.exebeche.com/ ITT: the iPad becomes a viable development platform.
01:45:52 <alise> I have but one thing to say:
01:46:08 <alise> well -- apart from the lack of a compiler :P
01:46:20 <alise> CakeProphet: Because, dude, heapsort just isn't hardcore enough!
01:52:05 <CakeProphet> so would now be the time to make a proprietary, better Erlang with C syntax?
01:52:28 <CakeProphet> With 100-core processors on the drawing boards these days.
01:54:15 <CakeProphet> I think if Google spends more time on Go that's probably what it will become, actually.
01:55:24 <alise> Go is not proprietary, and it is not Google's project.
01:55:39 <alise> Yes, Google advertised it, and it is worked on by Google employees, but actually more than anything it is a project of the Plan 9 development team.
01:58:23 -!- SevenInchBread has joined.
02:00:22 * alise tries to decide whether or not to use drop caps here
02:00:27 -!- CakeProphet has quit (Ping timeout: 248 seconds).
02:02:24 <alise> Gregor: Meet the old boss, same as the new boss.
02:05:27 <Sgeo> Does my source of opinions (alise) think Go is good?
02:05:46 <Gregor> OK, I've got a bunch of packages on floppies that are poorly labeled ... where's my damned 'cc' :P
02:06:07 <Sgeo> Well, that must mean it's incredible. If alise doesn't hate it...
02:06:44 <alise> Anyone want a ~super-high quality~ LaTeX-typesotten version of The Metamorphosis?
02:07:11 <Gregor> Apparently I'm installing "Operations, Administration and Maintenance"
02:07:16 <alise> It isn't so hot on screen; it alternates the margins of each page for printed use, and has blank pages so that chapters always start on the right page (I can't memorise recto and verso...)
02:07:19 <Gregor> That package name tells me just about zero :P
02:07:41 -!- Gracenotes has joined.
02:07:46 <alise> I used the Project Gutenberg translation, which isn't so bad.
02:08:34 <alise> It's set to A4 paper but I'm pretty sure it'd look too big; I'd just print it scaled down on something smaller.
02:09:47 <SevenInchBread> Is there a Haskell utility that automatically annotates functions with type declarations?
02:09:52 -!- SevenInchBread has changed nick to CakeProphet.
02:10:39 <Gregor> Surely this must come with a compiler X-D
02:10:53 <Gregor> I mean, maybe not? But probably!
02:11:25 <alise> Gregor: I imagine it comes with pcc.
02:11:47 <alise> The Portable C Compiler (also known as pcc or sometimes pccm - portable C compiler machine) is an early compiler for the C programming language written by Stephen C. Johnson of Bell Labs[1] in mid-1970s—based in part on ideas from earlier work by Alan Snyder in 1973.[2][3]
02:11:47 <alise> One of the first compilers that could easily be adapted to output code for different computer architectures, the compiler had a long life span. It shipped with BSD Unix until the release of 4.4BSD in 1994—when it was replaced by the GNU C Compiler. It was very influential in its day, so much so that at the beginning of the 1980s, the majority of C compilers were based on it.[4]
02:12:07 <Gregor> I'm installing package after package and not finding it.
02:12:16 <alise> For separate purchase, then.
02:13:08 <Gregor> That's probably true, but SO MUCH LAME.
02:13:10 -!- relet has quit (Quit: Leaving.).
02:13:11 <Gregor> I mean, come on, this is Unix!
02:13:16 <alise> "Skulpt is an entirely in-browser implementation of Python."
02:13:19 <alise> No... I ... can't believe that.
02:13:37 <ais523> can anyone here remember: which byte order is network order?
02:13:47 <alise> ais523: the one most processors use, I think... I may be wrong.
02:13:54 <alise> Gregor: And guess what? UNIX didn't include a C compiler until the 3rd Edition.
02:13:56 <ais523> Wikipedia was entirely unhelpful
02:13:58 <alise> ais523: *I have no idea*
02:14:07 <ais523> I have about half an idea
02:14:08 <alise> Gregor: It wasn't WRITTEN in C until the 4th Edition.
02:14:18 <alise> ais523: but ther are two options, so that's useless
02:14:22 <Gregor> alise: I realize that, but after that C and UNIX were inseparable.
02:14:25 <Gregor> alise: WHO THE FUCK KNOWS!
02:14:34 <ais523> Perl manual says big-endian
02:14:51 <alise> Gregor: Oh, of course; you'd buy the manual on paper, or it would be included with the system.
02:14:52 -!- Oranjer has left (?).
02:15:04 <alise> Did nroff even exist then? I think they just used troff to publish it.
02:15:17 <alise> man(1) is like the Encyclopedia Britannica on a CD-ROM.
02:15:36 <alise> Gregor: Ever used 1st Edition Unix?
02:15:48 <Gregor> I managed to make it run on an emulator once.
02:15:57 <Gregor> Friggin' awesome/terrible :P
02:16:01 <alise> http://code.google.com/p/unix-jun72/
02:16:06 <alise> Fixed-up OCR'd version; runs with SIMH.
02:16:07 <Gregor> This is about the earliest Unix that is totally still Unix though.
02:16:25 <alise> Gregor: The nice thing about the 1st Edition Manual is that it lists, along with procedure names, their PDP-11 assembly calling sequence.
02:16:35 <alise> But not emacs WHAT DO YOU THINK THAT MEANS
02:16:39 <alise> Also, blame Bill Joy.
02:18:02 <CakeProphet> I think I've figured out the most pointless project ever.
02:18:17 <alise> done, many times, utilised, many times.
02:18:37 <Gregor> Argh. idcc doesn't actually work >_<
02:18:58 <CakeProphet> in Python, actually. And then used to interpret the python interpreter to invoke the C interpreter.
02:19:24 <alise> [ehird@ping lib]$ ls -lh libuClibc-0.9.31.so
02:19:24 <alise> -rwxr-xr-x 1 ehird users 191K Jun 21 02:16 libuClibc-0.9.31.so
02:19:29 <alise> What did pikhq do to get it so big...
02:19:54 <alise> You're almost not funny.
02:19:59 <alise> wait, why is it SO
02:20:36 <alise> I got a Linux kernel down to ~480 KiB compressed
02:20:49 <alise> It doesn't support block devices but it *will* be able to create a usable in-RAM system from floppy.
02:20:51 <Gregor> alise: Does it support enough to be reasonably called a "Linux kernel"
02:21:15 <alise> Gregor: It has devices, it only has tmpfs but that's okay because initramfs can read the floppy into RAM,
02:21:18 <alise> It has TCP/IP support,
02:21:27 <alise> It ... has Ethernet device support?
02:21:45 <alise> It doesn't have much but I fully expect it to run that tiny X server thing and Dillo.
02:21:49 -!- poiuy_qwert has joined.
02:21:51 <CakeProphet> speaking of P=NP, don't you also win the prize if you solve P!=NP?
02:21:57 <alise> CakeProphet: of course.
02:22:01 <Gregor> Somehow this is rebuilding the kernel when there's new shit activated...
02:22:07 <CakeProphet> which do you think is more likely to be true?
02:22:13 <Gregor> That to me suggests it's actually building some C code, but that seems absurd, frankly.
02:24:06 <CakeProphet> My next language shall be neither compiled nor interpreted in any way. Instead it will be "realized" by a realizer program. A realizer takes source code as input and constructs a physical machine that perofmrs the semantics of that code.
02:24:58 <Gregor> So, you're inventing ... FPGAs.
02:25:13 <alise> Gregor: It probably is rebuilding the kernel.
02:25:31 <CakeProphet> Gregor: you just gave me a great bot idea, actually.
02:25:41 <CakeProphet> a bot that counts acronyms that occur on #esoteric in a given day.
02:26:46 <alise> Do you need pseudo-terminals for getty?
02:27:48 <CakeProphet> are there any stack-based physical machines out there?
02:30:35 <ais523> alise: probably not, getty uses real terminals
02:31:35 <alise> Flinix will use mingetty, probably.
02:31:38 <alise> CakeProphet: yes, see: forth machines
02:32:20 <Gregor> Maaaan. The lack of a C compiler is really harshing my buzz here.
02:32:22 <CakeProphet> alise: wait. see them where? Where, in the 1940s, would I go to look up such a thing? The library is closed right now.
02:34:32 <alise> How do you configure busybox to enable/disable commands?
02:34:49 <Gregor> make menuconfig or edit .config
02:35:02 <alise> How helpful. Now explain how menuconfig hasn't got those options.
02:35:57 <alise> Hmm, it simply appears to be... hiding things from me.
02:36:18 <alise> I wsa in a subdirectory, heh.
02:37:50 <alise> FUCK, I lost all myconfig.
02:38:22 <CakeProphet> hmmm... I'm trying to figure out what the advantage of using a register machine over a stack machine would be.
02:38:31 <CakeProphet> a stack machine architecture would be awesome.
02:38:58 <Gregor> Maaan, gcc supported this system, but good luck ever finding a binary, plus of course I don't have the development headers so I'm just boned.
02:39:17 <alise> WHEREAS I AM JUST STONED
02:39:39 <alise> (Isn't that deep?)
02:39:45 <alise> ((That's what she said!))
02:39:55 <CakeProphet> Though I'm not sure how many of you folks condone such things. Is anyone else a cannabis smoker? :o
02:39:59 <alise> Okay, what the fuck, menuconfig.
02:40:06 <alise> We're breaking up.
02:41:32 <alise> "Small, simple, evil." --BusyBox on ed
02:45:30 <alise> mount? How can I mount anything? There is no block device support!
02:45:42 <alise> login? init? Why do I need those? init is just a shell script, there are no users other than root!
02:45:58 <alise> It's so easy to configure a Linux installation if your constraints are dictated by the almighty 2 MiB floppy disk.
02:46:08 <alise> You just disable EVERYTHING.
02:48:42 <alise> Actually, me and pikhq are going to use KDrive... so we will have X11.
02:49:30 <CakeProphet> I make my own custom desktop environment on top of ncurses
02:51:04 <Gregor> Wow ... I just typed the pathname of a real program (which I don't have) in UNIX V/386 into Google, and got NO responses.
02:51:07 <Gregor> How is that even possible?
02:52:58 <CakeProphet> in the context of your time, I would say it has something to do with it being 2010.
02:53:11 * CakeProphet can communicate with the future only through IRC
02:53:24 <alise> I am a sentient Jacquard Loom.
02:55:34 <alise> flinix is going to be so awesome.
02:55:48 -!- augur has joined.
02:55:56 <CakeProphet> things were admittedly difficult when I time traveled to the 1920s. I had to hack together an IRC client from an electric typewriter, which meant I had to manually type in IRC commands.
02:56:53 <alise> Put a floppy in an old computer, reboot. Voila, 800x600 or 1024x768 256-colour X11 session opens with welcome text, a terminal, and a menu.
02:57:11 <alise> You can run a command in the terminal to start your Ethernet network and connect with DHCP,
02:57:28 <alise> then open the menu and you can use Dillo to browse the subset of the interwebs that hasn't upgraded to 2.0 yet.
02:57:39 <alise> You can irc via ircii or some other similar thing.
02:57:47 <alise> Sure, you can't mount any disks, but who really cares?
02:57:58 <alise> Oh, and did I mention? A package manager means that you can install any package that will fit in RAM from the interwebnets.
02:58:33 <alise> Unfortunately, you will be unable to save these packages, even to another floppy. The kernel HAS NO FILESYSTEM SUPPORT beyond tmpfs, or block device support.
02:58:41 <alise> Instead, initramfs loads the floppy into RAM as a tmpfs; job done.
02:58:51 <alise> ais523: not automatic, you might need to manually configure the network
02:59:03 <alise> and also, every byte counts; easier to write a quick briefing on how to start the network than an automagic script.
02:59:27 <alise> But, point is: Put in floppy. Boot up. Usable graphical environment. The web. IRC. WTF awesome.
02:59:54 <CakeProphet> could always do some kind of CLOUD FILESYSTEM OOOOOOH
03:00:32 <alise> ais523: And all this with a modern kernel, uClibc, and Busybox.
03:00:41 <alise> A modern kernel! One without block device support, sure, but...
03:00:58 <ais523> aren't floppy disks block devices?
03:01:15 <alise> ais523: Yes. But think about how a Linux boot floppy works. The kernel never reads the floppy.
03:01:27 <alise> The initramfs loads the floppy into RAM as a tmpfs, then hands over to the kernel, which sees it as a tmpfs.
03:01:36 <ais523> it's certainly an ingenious way to make the disk read-only
03:01:44 <alise> That's not intentional.
03:01:52 <ais523> hmm, what about kernel modules in the package manager
03:01:55 <alise> It's just that even with all of this removed, the kernel is ~480 KiB.
03:02:01 <alise> ais523: No kernel module support, heavens no!
03:02:03 <ais523> so if you really need to be able to read a disk, you could download block device support and use it
03:02:06 <alise> Oh, and that 480 KiB figure is LZMA compressed.
03:02:09 <alise> So yeah, the kernel is big.
03:02:11 <alise> Features are baaad.
03:02:19 <ais523> alise: oh, in that case, how are you going to support hardware?
03:02:27 <alise> ais523: I may end up enabling block device support and filesystems out of sheer weakness if userspace doesn't take up as much space as I expect.
03:02:31 <ais523> just assuming a VGA screen or whatever?
03:02:58 <alise> ais523: We're going to use the KDrive tiny X11 server (part of X.org now) with SVGA.
03:03:02 <alise> (We = me and pikhq.)
03:03:08 <alise> So, yeah, just assume SVGA.
03:03:11 <alise> It has a wide enough range of resolutions.
03:03:34 <ais523> and presumably you're only supporting wired networking
03:03:40 <ais523> to avoid needing a bunch of drivers for wireless cards
03:03:51 <ais523> even then, how are you going to select which networking drivers to use?
03:04:00 <ais523> the three most popular, or whatever?
03:04:13 <alise> ais523: Broadcom drivers. Nothing else. Works fine.
03:04:17 <alise> Most stuff is generic enough these days.
03:04:38 <alise> Ethernet, you can't really go wrong with Ethernet. Ethernet works nowadays, all the old tutorials saying "Oh, get a Broadcom card!" are really just FUD by now.
03:05:04 <ais523> I realised it worked fine, but I thought it was due to the breadth of available drivers, rather than them all being the same
03:06:02 <alise> There's actually very few Ethernet drivers to enable in the kernel.
03:06:07 <alise> Broadcom is the only one that would ring anyone's bell.
03:08:09 <alise> Also no USB or sound support. (Sound support requires block devices.)
03:08:49 -!- BeholdMyGlory has quit (Remote host closed the connection).
03:11:36 <alise> ais523: what other programs do you think would be useful?
03:11:53 <alise> ais523: how small can you get nethack?
03:12:01 <alise> we're going to be very short for space
03:12:04 <ais523> not sure, there's definitely a version designed to fit on a floppy
03:12:14 <ais523> which presumably implies that it takes up most of the floppy, so probably too large
03:12:26 <ais523> enough to access a package manager is probably fine
03:12:36 <ais523> also, try to put at least one TC program on there, like a shell or something
03:12:53 <alise> Busybox has a shell, so yes.
03:13:04 <alise> It will also have gzip and tar and the like.
03:13:10 <alise> ...I've included a floppy formatter but have just realised that'll be useless.
03:13:22 <alise> ais523: if I disable all the extensions and patches, surely nethack isn't so big?
03:13:26 <alise> What if I UPX the executable?
03:13:35 <alise> The whole system will be LZMA'd, anyway.
03:15:20 <alise> ais523: I'll probably include some tiny mail client.
03:15:55 <CakeProphet> irssi is a pretty sweet terminal IRC client.
03:16:02 <alise> ais523: Problem is, with a 480 KiB kernel, compressed, and the hundred kibibytes or so busybox will take up, we're short for space, even using an unformatted floppy disk.
03:16:06 <alise> CakeProphet: But it is too big for a floppy.
03:16:51 <alise> 2 mibibytes. Get a working system with X11.
03:17:26 <CakeProphet> the problem is that you're using /space/ to represent programs
03:17:36 <ais523> alise: can bootloaders read unformatted floppy disks?
03:17:47 <alise> I'm sure some can.
03:17:49 -!- FireFly has quit (Quit: swatted to death).
03:17:51 <alise> The formatting is just fluff.
03:18:07 <ais523> yes, but if they don't, there's no point in putting the disk in a computer and hoping it'll boot
03:18:10 <ais523> um, I don't mean bootloaders
03:18:36 <alise> Um, don't they just read the first sector and jump to it...?
03:21:42 -!- augur has quit (Ping timeout: 265 seconds).
03:22:42 <CakeProphet> hmmm... I've always wondered how LiveCDs mangage to be architecture independent.
03:22:48 -!- augur has joined.
03:23:11 <CakeProphet> don't they have to store some kind of machine code?
03:23:14 <alise> ais523: /init is just going to be a shell script that starts some gettys and X :-)
03:23:38 <ais523> CakeProphet: normally they use a machine-code polyglot
03:23:45 <ais523> not architecture-independent, but working on more than one arch as a result
03:24:06 <ais523> machine code is a pretty simple language to polyglot, generally, as it has no syntax errors
03:24:18 <ais523> it's like INTERCAL, only the bits you actually try to run have to make sense
03:24:19 <alise> most livecds aren't architecture independent
03:25:40 <CakeProphet> but when writing something like... a bootable floppy.
03:25:48 <CakeProphet> you'd want to imploy such a strategy right?
03:27:09 <CakeProphet> to run on any of numerous old CPU architectures of course.
03:28:07 <alise> how big is this hypothetical floppy?!
03:28:50 <CakeProphet> well, your bootloader is always the first sector, not the whole floppy. But I assume what you're saying is that such a thing is impractical.
03:28:54 -!- GreaseMonkey has joined.
03:29:26 -!- augur has quit (Ping timeout: 265 seconds).
03:31:12 <CakeProphet> what would be the most difficult combination of languages to polyglot. I guess imposing a limit of 2 languages.
03:31:36 <ais523> some are completely impossible, because they each require incompatible fixed headers
03:33:50 <CakeProphet> if only Erlang's compiler directives were -- instead of - :P
03:44:52 <CakeProphet> but I found found it. 5-language polyglot.
03:46:31 -!- alise has quit (Quit: Leaving).
04:18:39 -!- coppro has joined.
04:20:04 <CakeProphet> it would be cool to have a language that compiles to various polyglot styles.
04:20:43 <pikhq> CakeProphet: Erlang and Haskell? Sure. Just needs *literate* Haskell. ;)
04:21:06 <CakeProphet> ah. But that sounds like cheating to me. :P
04:21:48 <pikhq> Note that Literate Haskell is defined by the Haskell98 report; just a nice feature of the language is all.
04:24:12 -!- zzo38 has joined.
04:26:57 -!- cal153 has joined.
04:34:51 <zzo38> I have idea for a executable format in Linux that has something similar to COM format in DOS. But, instead, there is a small header in the file, to identify the file as this format and to tell how much memory to allocate, maybe also operating system and processor codes. Once the program is loaded, that header is replaced by the PSP in memory. And then the same memory is used for code/data/etc
04:36:02 <pikhq> That format is called a.out
04:38:48 <zzo38> No, a.out is different (I looked it up on Wikipedia)
04:40:51 <coppro> this conversation is full of fail
04:42:01 <zzo38> coppro: Please explain better then, if everyone else failed then maybe you can explain better?
04:42:07 <pikhq> Your mother is fail.
04:42:34 <coppro> zzo38: it's not that you failed. It's that you said things which are full of fail
04:43:16 <zzo38> Please describe, then, which things are full of failed
04:43:41 <coppro> for starters, your the fact that you don't understand the use of 'fail' as a noun
04:44:22 <zzo38> OK. Anything else?
04:45:57 <zzo38> Please be specific so that I can know what is wrong and so that I can fix it.
04:46:14 <zzo38> (The other thing you are fail is writing "else")
04:56:22 -!- augur has joined.
05:02:08 <CakeProphet> I've never quite understood how binary formats work.
05:02:27 <CakeProphet> Are they interpreted by something before they become machine code?
05:02:45 <pikhq> No, they are merely parsed for where to load stuff into memory.
05:02:52 <pikhq> They already *contain* machine code.
05:03:02 <coppro> the kernel (usually) reads them to set the program up
05:03:27 <CakeProphet> hmmm, okay. Got it now. I just never figured out where that happens I guess.
05:03:39 <pikhq> ELF just states "These places need to be replaced with addresses corresponding to these symbols" and "This needs to be loaded at place X".
05:03:50 <pikhq> Well. And a bunch of other metadata.
05:04:04 <CakeProphet> I assume these are things that machine code can't do alone?
05:04:19 <pikhq> It *is* machine code already.
05:04:53 <pikhq> Well, no. It's machine code and metadata.
05:05:02 <CakeProphet> ELF just states "These places need to be replaced with addresses corresponding to these symbols" and "This needs to be loaded at place X".
05:05:26 <pikhq> The machine code has a few variables it *might* need filled in; the metadata describes where those are and what they need to be filled in with.
05:05:44 <pikhq> (along with obvious stuff like "size of machine code", "requested load address", "CPU architecture", etc.)
05:06:15 <CakeProphet> what do you mean by variables? I'm not very familiar with machine code.
05:06:31 <pikhq> Some of these variables (usually called "symbols") might need a shared library; if so, the ELF file also describes which shared libraries are needed.
05:06:55 <pikhq> They're just memory addresses.
05:07:15 <coppro> don't they usually have a dynamic thunk?
05:07:17 <CakeProphet> so it contains machine code with macros (symbols), essentially?
05:07:45 <pikhq> coppro: Oh, right; it specifies the name of the dynamic linker too.
05:07:53 <pikhq> CakeProphet: For the most part.
05:08:32 <CakeProphet> because there are some things the compiler of said binary won't really know until the program executes?
05:09:11 <pikhq> Like which memory address functions from shared libraries happen to be loaded at.
05:09:57 -!- augur has quit (Ping timeout: 265 seconds).
05:17:01 <zzo38> I think you also need a simple executable format, the header needs only: Magic number; Processor type; Operating system type; Amount of memory to allocate (the same memory is shared for code/data/stack); Execute start address.
05:17:20 <zzo38> This is my idea of one kind of execute format
05:17:24 <zzo38> that can be invented
05:17:35 <pikhq> Having the stack anywhere near code/data makes me wince.
05:18:12 <coppro> just put it so that an overrun will hit you into code, which should be protected
05:19:52 <CakeProphet> well, fixing security risks from overflows really should be done at a library level. I don't think executables should worry about it.
05:20:00 <zzo38> If you allocate enough memory, the stack won't be near the code/data as much. But if you think your program might stack overflow, the program can reallocate the stack when it starts, anyways.
05:20:50 <coppro> most programs are not smart enough
05:22:21 <zzo38> This kind of executable format would be used mostly for if you are writing a small program in machine-codes which does not use a lot of stack space.
05:22:36 <zzo38> But can be used for larger programs as well.
05:23:45 <zzo38> Simply: The operating system allocates the memory, and then copies the contents of the file into the allocated memory, initialize stack pointer to the end of the allocated memory, replace the header with the PSP, and then jump to the execute start address.
05:25:45 <coppro> thank you for saying it, xckd
05:27:06 <fizzie> DOS EXE format is not too far from that, except it's got relocation entries too. But it does allocate a single block of memory (for code and data, plus a fixed amount of extra memory for zero-initialized data). And you specify the initial stack pointer in the header, so it can be put at the end there. (I don't quite know where it usually is.)
05:27:50 -!- augur has joined.
05:31:16 <Gregor> I'm desperately trying to make a cross-compiler targeting System V/386 :P
05:31:21 <Gregor> I'm betting: miserable failure.
05:31:54 <pikhq> Absolutely miserable, I'm sure.
05:32:17 -!- augur has quit (Ping timeout: 245 seconds).
05:32:29 <Gregor> newlib actually has a sysv i386 target.
05:32:38 <Gregor> I have no idea whether it's maintained or has a non-zero chance of working.
05:32:46 <Gregor> But since I don't have the native libc, it's all I've got :P
05:33:15 <Gregor> Or rather, I have the native libc, but not the headers and other stuff necessary for compilation.
05:34:37 -!- augur has joined.
05:36:03 <zzo38> fizzie: My idea is actually different, for one thing is not limited to DOS. For another thing the header is shorter, has no relocation entries, has no checksums or initial stack setting, no overlay numbers, and only one number for how much memory, which includes the entire file as part of it.
05:36:58 <fizzie> Yes. I just said "not too far".
05:38:43 <fizzie> I don't advocate using DOS EXEs, it's a pretty messy thing. The relocs are all segment-address fixups.
05:40:53 -!- augur has quit (Ping timeout: 265 seconds).
05:44:25 <zzo38> I don't want to use a complicated bootloaders such as GRUB and so on, I want to use a simple one! I wrote a thirty-six bytes MBR code, so it should be mostly good enough.
05:50:20 <Gregor> Daaaaamn GCC takes forever to compile.
05:50:48 <Gregor> And all of this is just going to result in a nonworking GCC anyway :P
05:52:00 -!- CakeProphet has quit (Ping timeout: 265 seconds).
05:53:08 <pikhq> GCC takes nearly forever to compile.
05:53:18 <pikhq> zzo38: Does it load Linux?
06:06:43 -!- augur has joined.
06:06:44 -!- augur has quit (Remote host closed the connection).
06:07:21 -!- augur has joined.
06:07:37 <zzo38> pikhq: The thirty-six bytes MBR code does not load Linux. It loads only real-mode programs, and not more than 64K. However, it can be used to load a second loader, which then initializes protected mode and loads Linux.
06:09:07 <pikhq> So, it loads a COM file.
06:09:39 <zzo38> pikhq: Yes, except that there is no PSP and no DOS functions.
06:11:50 <zzo38> I don't think a code to fully load Linux (or even to fully initialize protected mode properly) will fit in the MBR code.
06:12:55 <pikhq> Though to load Linux does *not* entail protected mode.
06:13:30 <pikhq> It entails loading Linux and an initrd into memory with a command line at a given place, and jumping into its initialisation routine.
06:14:03 <zzo38> And how long is a code to do that?
06:15:13 <pikhq> 512 bytes, I think.
06:15:37 <pikhq> I seem to recall Linux 2.4 had such a bootloader as its first 512 bytes.
06:15:52 <zzo38> That won't fit in the MBR code. the MBR is 512 bytes long but part of it is used for magic number and stuff
06:16:24 <fizzie> There is (or was, maybe) a boot sector in the kernel, yes, but it was pretty limited. Didn't do any filesystems, for example, so it could only load the kernel from a raw media.
06:16:45 <fizzie> And it didn't do the usual sort of separate-file initrd, it did something else instead.
06:17:15 <fizzie> It did support flipping floppies in-between loading the kernel and the ramdisk, though.
06:17:21 <pikhq> You could specify how many blocks in the initrd was using some tool to set the default command line.
06:17:28 <pikhq> Or just have it swap floppies.
06:17:43 <pikhq> It *was* quite nice to do dd if=bzimage of=/dev/fdd, though.n
06:23:45 <Gregor> Holy fekk GCC compiled!
06:26:06 <fizzie> pikhq: In arch/x86/boot/header.S of 2.6.33.1 (for which I had sources handy) there is in fact still a boot sector, but the boot sector just prints out the text at address "bugger_off_msg", namely "Direct booting from floppy is no longer supported. Please use a boot loader program instead."
06:26:21 <pikhq> Ah, right. Gah I hate that.
06:26:48 <pikhq> I can't imagine keeping around a freaking boot sector would take too much work.
06:27:15 <pikhq> It's not like the BIOS has changed since, oh, DOS 1.
06:31:26 -!- tombom has joined.
06:33:25 -!- CakeProphet has joined.
06:36:35 <Sgeo> My dad is an idiot.
07:00:25 -!- coppro has quit (Ping timeout: 264 seconds).
07:02:33 -!- SevenInchBread has joined.
07:05:09 -!- CakeProphet has quit (Ping timeout: 240 seconds).
07:11:20 <zzo38> How often do you use dual-licensing?
07:11:37 <ais523> I more often use a general licence, than two licences
07:11:47 <ais523> e.g. release something LGPL even though I only plan to use it as GPL
07:12:13 <Gregor> MEMDISK is insufficient to boot UNIX SysV.
07:12:30 <Gregor> Which means unfortunately that UNIX SysV can't be turned into a delightfully absurd LiveCD.
07:12:54 <zzo38> ais523: Yes I can understand that, which I might do in some cases (but mostly not). Usually I don't use dual-licensing either. But sometimes I do.
07:13:21 <pikhq> Sgeo: Yes, your father is the embodiment of idiocy.
07:13:22 <zzo38> I wrote one program which is dual-licensed by "GNU GPL version 2 or later version, or Ms-RL (Microsoft Reciprocal License)".
07:13:53 <zzo38> Only one program, though. The reason why I did that is because it is a program which would be useful for Microsoft to include in Windows.
07:14:11 <zzo38> I would also be useful to include in ReactOS.
07:14:41 <zzo38> This particular program happens to be useless for Linux, however.
07:15:04 <ais523> zzo38: somehow I doubt Microsoft would include it in Windows even if it were MS-RL licensed
07:15:09 <ais523> they're rather paranoid
07:15:49 <Sgeo> He doesn't want "other people on our network"
07:16:26 <Sgeo> Best explanation he can give for his "no hosting" rule
07:16:28 <zzo38> ais523: You are probably right, but they can do so if they want to.
07:16:55 <Sgeo> And that he needs the bandwidth. And saying "Ok, no bandwidth-intensive stuff when you're home" isn't enough.
07:17:49 <ais523> he needs upload bandwidth?
07:18:01 <ais523> hosting mostly uses a different sort of bandwidth from normal internet use
07:18:43 <Sgeo> I.. don't think so. I wasn't even thinking of that, but he needs mostly download bandwidth, but this was apparently a separate issue
07:19:39 <zzo38> I host three protocols on my computer, and my father is OK with that. (And later I might even add more protocols)
07:20:11 <Sgeo> My dad won't even let me host games
07:20:28 <Sgeo> And he won't let me run a client that runs a game if I'm not attending it
07:20:47 <Sgeo> Actually, I think he wants me to cut off Internet access when I'm not at the computer
07:21:23 <fizzie> Most ADSL-providing ISPs around here have slipped a "you must not connect any sort of servers to your pipe" condition to their terms and conditions, though that gets widely ignored (esp. for games), I think.
07:21:23 <pikhq> I think he thinks that bandwidth is a *resource*.
07:21:34 <pikhq> Rather than, y'know, a measurement of the size of your pipe.
07:21:56 <fizzie> If you eat all your bandwidth cake today, you'll go hungry tomorrow!
07:22:02 <Sgeo> I asked if there was a cap, he said no
07:22:24 <Sgeo> Anyways, watching the last 10 min of this SGA ep
07:22:25 <pikhq> Okay, so he knows nothing about hor the Internet works.
07:22:37 <pikhq> Clearly, he should head a Congressional commitee on it.
07:23:04 <Sgeo> He once made me wipe my HD because I downloaded and installed a chat server
07:23:41 <zzo38> I know the ISP that I use is OK with hosting any protocols (I even asked them).
07:24:16 <pikhq> He is too stupid to be allowed to own a computer.
07:24:34 <pikhq> No, he is too stupid to be allowed to own any technology more complex than flint & steel.
07:25:47 <zzo38> pikhq: You might be right (approximately)
07:30:19 -!- pikhq has quit (Remote host closed the connection).
07:50:31 -!- tombom has quit (Quit: Leaving).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:14:39 <ais523> fizzie: ironically, there are ISPs that actually work like that
08:26:26 -!- zzo38 has quit (Remote host closed the connection).
09:15:22 -!- FireFly has joined.
09:19:04 -!- MigoMipo has joined.
09:22:19 <ais523> any File Hierarchy Standard experts here? I'm trying to figure out where you're supposed to put architecture-independent binaries not meant to be run by the user
09:22:26 <ais523> which is a possibility I don't think they considered
09:28:51 <ais523> my current example is a Java binary
09:28:58 <ais523> which runs on the JVM, thus platform-independent
09:31:52 <Ilari> I would stick those (but then, I have pretty much no clue of FHS) in share directory.
09:32:11 <ais523> atm I'm putting it in a subdir under usr/lib
09:32:22 <ais523> which is at least vaguely consistent with the other rules
09:32:33 <ais523> it would obviously be the right place for the file if it was written in, say, C
10:08:25 -!- MigoMipo_ has joined.
10:12:16 -!- MigoMipo has quit (Ping timeout: 260 seconds).
10:13:31 -!- MigoMipo_ has changed nick to MigoMipo.
10:14:01 -!- MigoMipo has quit (Remote host closed the connection).
10:18:28 -!- SevenInchBread has changed nick to CakeProphet.
10:28:09 -!- dbc has joined.
10:28:49 -!- atrapado has joined.
10:30:47 -!- GreaseMonkey has quit (Quit: I'm using NO SCRIPT WHATSOEVER - Download it at file:///dev/null).
10:36:17 <fizzie> ais523: FWIW, Debian puts "library"-ish Java binaries in /usr/share/java/.
10:36:58 <ais523> ah, I may move mine there then
10:37:13 <ais523> what if they're only meant to be run by one particular shellscript?
10:38:50 <fizzie> I'd say there in that case too. At least ant is there, and it's mostly meant to be run by the /usr/bin/ant script, though admittedly IDEs and other such things use the Java classes "directly" too.
10:39:40 <fizzie> After all, with Java, you never quite know who'll want to reuse bits and pieces of you.
10:40:32 <fizzie> On the other hand, Eclipse puts all its Java stuff into /usr/lib/eclipse/.
10:40:50 <fizzie> (Well, all its stuff in general, I guess.)
10:41:31 <fizzie> But OpenOffice.org's Java bits are in /usr/share/java/openoffice/ even though that's quite many files.
10:41:44 <fizzie> It doesn't seem to be much of a consensus.
10:55:20 -!- MigoMipo has joined.
11:00:40 -!- augur has quit (Remote host closed the connection).
11:08:49 -!- bsmntbombdood has quit (Ping timeout: 276 seconds).
11:15:10 -!- oerjan has joined.
11:34:58 <Ilari> Aren't there 32- and 64-bit versions of Java (differing mainly in array indexing). Are the bytecode files cross-compatible (32-bit ones just can't access array indexes beyond 2Gi)?
11:35:22 <ais523> yes, not only that, but the bytecode format AFAICT has never been changed
11:35:45 <ais523> you'll get class-loading errors if you try to run a recent Java file on an old JVM and you reference standard classes that used not to exist, but that's about it
11:35:57 <Ilari> Lua bytecode files are arch-specific...
11:35:58 <ais523> (and you can use reflection to get around that problem)
11:36:10 <ais523> the JVM arch-portability is the whole point
11:37:05 <Ilari> Heh... Mention of errors bring me flashbacks from making app that has optional GUI and from making app that uses JNI and works on both GCJ and Sun JVM...
11:37:33 <Ilari> (two different apps, both wound up having to catch errors)...
11:40:21 <Ilari> GCJ links JNI libs into executable, so you don't have to load them and attempts to load them would error out anyway. Sun JVM requires loading the libs...
11:41:06 <Ilari> And if you call GUI function when there is no GUI available (and not in headless mode) you get an error (not merely an exception).
11:41:22 <Ilari> (some class does not exist error).
11:50:47 <fizzie> At least you can be assured that whatever is thrown is a subclass of Throwable. In C++ someone could throw an elephant at you.
11:51:38 <fizzie> I've seen code that throws undocumented unsigned integers, and has a single top-level catch which prints "error n" and dies. Helpful!
11:52:42 <fizzie> And multiple instances of code that throws string constants that are used as the error messages. (Which is pretty Perlish in a way.)
11:54:51 <fizzie> I'm tempted to try out what happens if I throw a non-Throwable from a JNI method, but this phone keyboard is not very comfortable for coding.
11:55:02 <ais523> fizzie: I remember when I was learning C++, I threw integers just because I could
11:55:15 <Ilari> (Well, error can take any type of value, but usually it is string that is "thrown").
11:56:45 <Ilari> Apparently in one piece of code, due to bugs threw std::string's...
11:58:14 <Ilari> (just guess what kind of programming error causes code that should throw subclasses of std::exception to throw std::string's instead...)
11:58:32 <ais523> throwing the message of an exception rather than the exception itself?
12:00:21 <Ilari> Pretty much. Forgetting to actually specify the class to throw ('+ std::string +' makes it std::string).
12:00:36 <Ilari> Like 'throw ("foo" + bar)'.
12:00:58 <Ilari> When it should be 'throw std::runtime_error("foo" + bar)'.
12:01:16 <AnMaster> ais523, what is wrong with being able to throw integer?
12:01:45 <ais523> AnMaster: it becomes unmaintainable pretty easily
12:01:54 <ais523> and interferes with the use of exceptions for exception handling
12:02:10 <AnMaster> ais523, how would it work in languages without objects but with exceptions?
12:02:15 <ais523> if the exception is unhandled, whatever code is calling you, rather than getting a nice exception object it can query and show to the use, gets a meaningless number
12:02:17 <AnMaster> if you weren't allowed to throw any data type
12:02:27 <ais523> I see nothing wrong with allowing it
12:02:34 <ais523> but even if allowed, I don't think it's a feature that should be use
12:02:52 <AnMaster> ais523, I suggest throwing meaningful tuples instead
12:02:58 <Ilari> At least in Lua, one can "throw" a table...
12:03:30 <ais523> AnMaster: is that meaningful?
12:04:07 <AnMaster> erlang has exceptions and allows you to throw pretty much any data type (iirc there is some restrictions wrt "references" to things like open files and such)
12:04:38 <AnMaster> and erlang doesn't have object orientation so restricting it to "exceptions" wouldn't work
12:04:50 <Ilari> Worst is where code throws integers that change (new important ones are added) between versions. And there's no integer to string method...
12:04:53 <AnMaster> usually you throw tuples or atoms
12:05:03 <AnMaster> and document what stuff you may throw
12:05:33 <AnMaster> when you get a runtime exception it tends to be a tuple with the first member being an atom indicating the type
12:05:39 <Ilari> So caller may suddenly get "Error 75" and has no idea what it is supposed to be...
12:06:01 <AnMaster> such as badmatch, or badarith (usually div by zero)
12:07:19 <fizzie> Perl does allow throwing references in addition to just strings, I'm not sure how much that gets used though.
12:08:13 <AnMaster> fizzie, I think you can throw some references, such as pids. What I'm unsure about is if you can throw ports or not. ports is a low level stuff in erlang which is basically a thin wrapper around an fd most of the time
12:08:21 <AnMaster> usually you don't work directly with ports anyway
12:09:33 <ais523> ooh, now I want to write a Perl program that throws a typeglob
12:09:35 <fizzie> Perl "references" is perhaps a bit different term; usually it's just references to arrays or hash tables.
12:09:38 <ais523> but I'm not entirely sure why
12:09:59 <fizzie> ais523: I was just about to say I wouldn't try that. :p
12:10:11 <AnMaster> hm erlang has references too. In that case it is a kind of "guarantied unique id" thingy
12:10:15 <ais523> Perl references are like C pointers, except they're refcounted and you can't do arithmetic on them
12:10:50 <Ilari> What's a typeglob?
12:10:53 <fizzie> You could probably throw an IO::Handle without much problems in Perl, but it's not as distributedy as Erlang.
12:11:12 <AnMaster> basically there is a runtime function that will return a new reference every time, and you can't construct arbitrary references out of the air. Well perhaps you can with some debugger function, wouldn't surprise me.
12:11:36 <AnMaster> fizzie, oh yeah ports are local to the node. Which is why you don't work with ports directly most of the time
12:11:56 <AnMaster> you work with an IO server process provided by the runtime by default unless you open the file in a raw mode.
12:12:18 <AnMaster> well there are other good reasons for the io server process. Like it provides buffering.
12:12:49 <ais523> Ilari: don't ask what a typeglob is, it takes too long to explain and even when you do you get it wrong
12:13:02 <AnMaster> fizzie, I assume you could throw a fun *tests*
12:13:03 <ais523> IIRC most explanations of how they work basically just give you a list of examples
12:13:12 <AnMaster> fun being like a lambda or a function pointer
12:13:42 <fizzie> http://www.sdsc.edu/~moreland/courses/IntroPerl/docs/manual/pod/perldata.html#Typeglobs_and_Filehandles has (one) official explanation.
12:13:47 <AnMaster> 2> catch throw(fun erlang:abs/1).
12:13:56 <fizzie> (Or your local perldoc perldata.)
12:14:14 <AnMaster> that is function pointer style
12:14:16 <AnMaster> 3> catch throw(fun(X) -> X*X end).
12:14:20 <AnMaster> lambda style works just as well
12:16:20 <AnMaster> Returns an almost unique reference.
12:16:20 <AnMaster> The returned reference will re-occur after approximately 2^82 calls; therefore it is unique enough for practical purposes.
12:16:35 <fizzie> throw(fun party/2) -- when you want to arrange a pleasant party for two.
12:17:33 <AnMaster> fizzie, well that requires a function party in the same module that takes two parameters
12:20:58 <AnMaster> 6> (catch throw(fun(X) -> X*X end))(3).
12:21:36 <AnMaster> (catch throw(fun party/2))(self(),
12:22:48 <AnMaster> (self() is a function that returns the pid [that is erlang pid, not related to OS pids at all] of the current process btw)
12:27:37 -!- oerjan has quit (Quit: Arrivederci).
13:20:20 -!- Gracenotes has quit (Ping timeout: 240 seconds).
13:23:45 -!- augur has joined.
13:24:35 -!- Gracenotes has joined.
13:35:58 <oklopol> "<Sgeo> He doesn't want "other people on our network" <<< maybe he downloads a lot of CP and is paranoid because of that?
13:36:23 <oklopol> (thought i typoed and deleted the other one)
13:52:38 <fizzie> Hrm, an ad banner with the link URL ending ...&cb=INSERT_RANDOM_NUMBER_HERE. I guess they assume a DWIM-browser.
13:53:27 <ais523> or DWIM server to serve the ads
13:53:39 -!- BeholdMyGlory has joined.
14:04:18 <AnMaster> there are some speed reduction thingies on this street, in the form of two boxes that forces you to slow down and drive in the shape of an S
14:04:41 <AnMaster> today, two different lorries have drove up to it, and then backed down the whole street
14:05:40 <AnMaster> and just now a tow truck doing the same
15:43:47 -!- lifthras1ir has quit (Remote host closed the connection).
15:58:24 -!- lifthrasiir has joined.
15:58:44 -!- cpressey has joined.
15:59:14 <AnMaster> how strange for an embedded system to have a snprintf() that inserts "(null string)" if %s is passed a NULL pointer
15:59:24 <fizzie> cpressey: A real .com file from an unmodified Linux x86-64 GCC+ld, with command line arguments only (no fiddling with .specs): http://pastebin.com/eChGDKDy
15:59:27 <AnMaster> I mean, I would have expected undefined behaviour
16:00:01 <AnMaster> fizzie, btw I managed to get some more free space on that device by tuning off features in config.h that I didn't use
16:00:21 <fizzie> AnMaster: Well, if it's undefined, "(null string)" is perfectly valid too.
16:00:54 <AnMaster> fizzie, yes but you wouldn't expect an embedded OS to waste space on 1) the code to check for that 2) the space to store the constant "(null string)"
16:01:50 <fizzie> On the other hand, it's more robust that way.
16:02:17 <AnMaster> but since this doesn't have an MMU there are loads of other ways to crash it anyway.
16:02:32 <AnMaster> battery.o: DBase 3 data file with memo(s) (813309772 records)
16:03:13 <AnMaster> objdump is way more correct: battery.o: file format coff-h8300
16:04:03 -!- pikhq has joined.
16:05:05 -!- Phantom_Hoover has joined.
16:07:31 <AnMaster> fizzie, a lot of linker fiddling though it seems
16:07:34 <fizzie> cpressey: It works in dosemu, but I'm a bit suspicious of how I'm accessing 0xa0000 via ds initialized to 0; I thought in real mode it would automatically set the segment limits to 64k. (It's easily fixable by putting 0xa000 into ds and initializing the videomem ptr to 0 instead.) And it's ugly code since GCC generates 32-bit everything, and ".code16gcc" just adds instruction prefix bytes to make that work in a 16-bit segment.
16:07:55 -!- lifthrasiir has quit (Remote host closed the connection).
16:08:55 -!- lifthrasiir has joined.
16:09:12 <fizzie> AnMaster: Yes, and I couldn't quite figure out how to make it put the .data segment immediately after .text (without a separate linker script file); done that way it'll probably lose if you put in any (initialized or uninitialized) static data, except if you put in a __attribute__ ((section (".text"))) in.
16:09:36 <fizzie> Phantom_Hoover: No, I just started when cpressey came in.
16:09:53 <fizzie> Phantom_Hoover: Good timing there.
16:11:07 <pikhq> fizzie: dosemu is actually running in a psuedo-real mode. If it works there, it *should* work on actual DOS.
16:11:21 <fizzie> pikhq: Uh, sorry, dosbox is what I meant to say.
16:11:23 <cpressey> 17? There was a weekend in between... we've been talking about it for 65 hours straight!
16:11:47 <pikhq> Oh. DOSbox *ought* to work, as that's actual (emulated) realmode, but there is a chance of bugs.
16:12:07 <fizzie> Right; I'm not sure how thorough they are with sanity-checks against illegal behaviour like that.
16:12:17 <AnMaster> also who knows if it would work with different drivers and such loaded
16:13:47 <fizzie> Dosbox has a couple of different CPU core emulation modes, especially the dynamic-recompilation one might be a bit fast-and-furious here and there.
16:15:07 <AnMaster> how do you list sections for a coff file? readelf -S obviously doesn't work
16:15:31 <AnMaster> I have binutils tools for the architecture available, not much else
16:15:46 <pikhq> Get a binutils for Windows and objdump?n
16:15:49 <fizzie> Well, "objdump -h" ("section headers") or something.
16:15:58 <AnMaster> pikhq, what would that help? it is not for x86 at all
16:16:23 <AnMaster> fizzie, yep that works. thanks
16:16:25 <pikhq> Which is nearly COFF.
16:16:34 <pikhq> But yeah; objdump for the relevant architecture.
16:16:50 <AnMaster> pikhq, a 16-bit platform. big endian too
16:18:28 -!- ais523 has quit (Remote host closed the connection).
16:25:59 -!- cheater99 has quit (Quit: Leaving).
16:28:15 <fizzie> AnMaster: Hopefully this'll mean I can now read that SDHC card too without going through the camera: http://sprunge.us/UdMB (and what a generic name there).
16:30:10 <fizzie> (On the other hand, the old card reader's ID for the different slots was actually "Generic [cardtype]".
16:33:47 <fizzie> We talked about card readers not many days ago.
16:33:54 <AnMaster> fizzie, as long as it is usb mass storage interface
16:34:10 <AnMaster> and not some other one. Let me find that hilarious driver in the linux kernel
16:34:52 <fizzie> The old one didn't do SDHC, just plain old-fashioned SD cards. (And I think it was the slower USB 1.1, this should be 2.0-capable. Not that my cards are probably very snappy either.)
16:37:44 <AnMaster> fizzie, found with google: http://lxr.free-electrons.com/source/drivers/mmc/host/ricoh_mmc.c?v=2.6.28 but it seems gone in newer kernels
16:37:59 <AnMaster> fizzie, read the comment starting with "This is a conceptually ridiculous driver"
16:38:05 <fizzie> Initial impressions; they've gone a bit overkill with the "I'm doing something" LED's brightness; it almost put my eye out. (Well, maybe not; but it has the "gritty" bloom-like effect I've only seen with laser pointers so far. Yes, yes, I know you're not supposed to look at those.)
16:40:34 -!- JodaZ_ has quit.
16:42:39 <AnMaster> http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.34.y.git;a=commitdiff;h=03cd8f7ebe0cbef5ca7eed349774085e92a3d726
16:45:08 <fizzie> USB devices are full of quirks. The PS3 "sixaxis" controller follows the USB HID spec in everything that matters, except that you need to send a HID_REQ_GET_REPORT request with a specific, nonstandard type (0xf2) before it goes into "operational mode" (i.e. works).
16:45:32 <fizzie> Personally I can't figure out any other reason for that except that they don't want people to be able to just plug that thing into a PC and have it working.
16:45:35 <cpressey> That's what makes it "Universal".
16:46:56 <AnMaster> fizzie, and sometimes the quirks mess up for other devices. Like this bug that was finally resolved in 2.6.34, had to revert that quirk for other products locally for quite a few releases... https://bugzilla.kernel.org/show_bug.cgi?id=14049
16:49:56 <AnMaster> who knows how long it will stay working for me however
16:52:10 <AnMaster> well, it's supposed to help haven't tested it yet
16:53:51 <AnMaster> I presume there is some init script
16:54:37 <AnMaster> Phantom_Hoover, afaik flightgear uses openal
16:54:50 <AnMaster> so I guess openal has issues with pa
16:55:12 <Phantom_Hoover> All I know is that in the past it has run much better without PA.
16:55:12 <oklopol> you know what's interesting
16:55:38 <AnMaster> Phantom_Hoover, ubuntu at least has /etc/init.d/pulseaudio
16:55:48 <fizzie> There's also "pasuspender" in many systems.
16:56:01 <fizzie> It does a "temporarily disable pulseaudio until a process has finished" thing.
16:56:07 <oklopol> if i make out with chicks who know i have a gf, they are fine with it without exception, once i mention i'm in an open relationship, they go all wtf get some morals dude
16:56:30 <fizzie> So you can do "pasuspender -- foo bar baz", and it'll run "foo bar baz", wait until it terminates, then restore pulseaudio functionalities.
16:56:34 <fizzie> I've never tried it, though.
16:56:56 <fizzie> It doesn't exactly shut down the daemon, it just tells it to disconnect all access to audio devices.
16:58:41 <AnMaster> how does one edit keyboard shortcuts in gnome for some application. gnome-terminal in this case.
16:58:54 <AnMaster> I thought there was some option to enable editing in the menus but I can't find it
16:58:58 <oklopol> so what's the puzzle today, oerjan
16:59:57 <oklopol> also people like to preach how that kind of relationships don't work without ever having been in as long a relationship as us
17:00:06 <AnMaster> Phantom_Hoover, I think it used to be, maybe they removed it -_-
17:00:18 <oklopol> yeah you talk about daemons boring people
17:00:26 <oklopol> or wait what are you talking about
17:00:49 <AnMaster> Phantom_Hoover, hm tell me if you find it
17:00:56 <AnMaster> Phantom_Hoover, I'm going to try gconf now
17:01:03 <oklopol> okay i can't tell, what's the topic
17:01:17 <AnMaster> Phantom_Hoover, exactly my feeling
17:01:22 <AnMaster> oklopol, are you talking to us?
17:01:38 <oklopol> no i'm asking myself what i'm talking aboutz
17:01:54 <Phantom_Hoover> oklopol, yes, people are weird about sexual mores. We get it.
17:02:03 <oklopol> Phantom_Hoover: i don't actually want comments
17:02:15 <AnMaster> Phantom_Hoover, found it in gconf anyway
17:02:27 <oklopol> i just like changing these kinds of topics, what topic is this?
17:03:09 <AnMaster> Phantom_Hoover, I'm just confused I can't find it in settings either on ubuntu or arch linux. But last I saw it was in jaunty
17:03:16 <AnMaster> and karmic I just ran for about 2 days
17:03:36 <oklopol> and i guess guis specifically suck
17:03:43 -!- MizardX has quit (Ping timeout: 276 seconds).
17:04:12 <fizzie> Phantom_Hoover: "pacmd exit" kills the daemon, but at least in my case it automagically starts up again all the time, possibly via dbus or who-knows-what-magic.
17:04:17 <AnMaster> Phantom_Hoover, I almost want to go to gnome's irc server and ask, but you never get any reply there
17:04:57 <fizzie> Possibly "sudo service pulseaudio stop" (or the likely-equivalent /etc/init.d/pulseaudio stop) will actually really make it stop.
17:05:01 <AnMaster> fizzie, heck I remember I used that back during KDE 3.x to prevent artsd from running
17:05:17 <oklopol> so how about this esolang that's like brainfuck but you have different kinds of fart sounds as commands?
17:05:32 <oklopol> one command could play a fart sound on the speakers
17:05:54 <AnMaster> #ifndef DOXYGEN_SHOULD_SKIP_INTERNALS
17:06:00 <oklopol> what are the fart sounds in english i wonder
17:06:02 <fizzie> Still, pasuspender (or just a manual "pacmd suspend") should make it detach and close the audio devices, so that they're available to other apps.
17:06:09 <oklopol> so we can get this thing started
17:06:21 <oklopol> probably best idea this year so i'm expecting full cooperation
17:07:10 <AnMaster> whatever happened to dmix or whatever the name was
17:07:17 <AnMaster> software mixer in alsa itself wasn't it?
17:07:25 <AnMaster> why would anyone need anything more than that
17:07:37 <AnMaster> I'm lucky my desktop has hardware mixer however
17:08:35 <fizzie> Many people need "more than that", but of course there was already jack and such for the audio-enthusiasts. Still, I don't feel pulse's so horrible; of course I haven't been bitten by very bad bugs (yet).
17:09:13 <AnMaster> fizzie, yes there is jackd for the "more than that" case. I don't see any niche for pulseaudio though
17:10:37 <fizzie> pavucontrol's reasonably nice; you see a list of all apps that are generating sounds, and can switch which pipe (digital output to the amp, USB sound card to headphones, bluetooth headset) you want them to make noises to, easily. Of course with most apps you can specify the ALSA device, but it might not be in a easily accessible place, hidden in some preferences dialog or other.
17:11:10 <oklopol> mostly i find "pfft"as a suggestion 8\
17:11:16 <AnMaster> fizzie, but who would play two different videos at once while also listening to music?
17:11:25 <AnMaster> fizzie, that seems like a non-existent use case
17:11:38 <oklopol> maybe finnish fart sounds then
17:11:50 <AnMaster> sure, pulseaudio fills it, but most such apps can set their own volumes
17:12:02 <AnMaster> not sure about mplayer, but vlc can and iirc so can xine
17:12:03 <oklopol> wtf, why wouldn't there be decent onomatopoeia for farts
17:12:16 <fizzie> AnMaster: Not at the same time (very often), but I do keep using e.g. different outputs for Flash sound, and for Flash it's really inconvenient to start changing the ALSA device it uses.
17:12:28 -!- MigoMipo_ has joined.
17:12:44 <oklopol> but anyway maybe we can use pf^+t for now do you agree?
17:12:55 <oklopol> pft and pfft could be + and -
17:13:06 <AnMaster> fizzie, file bug report for flash ;P
17:13:20 <oklopol> pffft isn't <, because that wouldn't be very esoteric would it?!? so we SKIP the three xD
17:13:42 <oklopol> and then maybe pffffft is > because on the other hand we don't wanna be TOO esoteric
17:13:51 <oklopol> okay getting kind of long, we'll carry to g
17:13:52 <AnMaster> fizzie, anyway, most cases of different volumes at the same time I can think of would be bg music/other sounds in games. And that would probably be mixed by openal before it reached alsa or pulseaudio
17:14:18 <AnMaster> oklopol, this is just like ook, rather boring
17:14:18 <oklopol> and outputs, well we just need the actual fart command, that'll be prrrrt no matter whether it's actually used, i like it.
17:14:30 <oklopol> no i think it's hilarious xxxxxxxD
17:14:47 -!- MigoMipo has quit (Ping timeout: 260 seconds).
17:14:48 <oklopol> we have much to talk about
17:14:56 <oklopol> and someone needs to start working on the spec
17:15:09 <fizzie> AnMaster: I don't really need the mixing-together part of pulseaudio (or ALSA's dmix, or a hardware mixer) very often.
17:15:12 <oklopol> does anyone have a weekend to spare soon?
17:15:35 <oklopol> this is a pretty important project so you can probably skip weddings and funerals and the like.
17:16:32 <oklopol> okay now to add esotericness, maybe just MAYBE there could also be, occasionally, fart sounds even if you don't use the command, and kinda like in intercal, if you don't fart often enough then the program crashes=!?=!?!? crashes = makes a fart sound and dies
17:17:28 <fizzie> oklopol: You can't find others to fart with?
17:17:40 <oklopol> you can fart in a microphone
17:17:47 <oklopol> and it interprets it somehow
17:17:54 <oklopol> obv you can just store farts on the tape
17:18:05 <oklopol> there are 256 different fart values
17:18:06 <fizzie> Just don't ask me to start building some sort of fart-recognition system.
17:18:27 <oklopol> well you're the sound recognition dude, so i was just assuming
17:18:42 <oklopol> anyway 0 is called the SMELLY FART
17:19:34 <oklopol> and i don't really understand how [ and ] work in brainf*ck we should probably copy the sementics
17:19:37 <fizzie> Do you have some sort of a "thing" with farts?
17:20:29 <oklopol> whether we have a fart fetish?
17:20:44 <oklopol> i should probably not irc anymore today
17:20:51 <fizzie> Yes, that might be safer.
17:21:26 <oklopol> i was gonna do stuff today
17:21:59 * oklopol considers putting "FARTS LOL xD" on the wiki
17:22:29 <oklopol> or F-101-xD, that looks pretty scientific
17:22:33 <fizzie> You've just been farting around, then, eh?
17:22:45 <oklopol> it's kinda like acronym but there's leet and cool smileys too
17:22:46 <fizzie> "The act or process of wandering aimlessly with no particular goal", according to ud.
17:23:07 <oklopol> i'm not sure i've been wandering
17:23:15 <oklopol> but i've been reading about nonwandering, does that count
17:23:18 <fizzie> "delaying something unnecessarily by dawdling, and slacking off"?
17:25:48 <fizzie> AnMaster: Oh, incidentally; if you remember the "gift card to a store that had closed down all their retail places" thing, I finally got an answer out of them (24-hour email response time my... well, backside); the gift card is unusable in their webshop, but works directly in those "Gigantti" retail stores; the systems are compatibbel.
17:26:41 <oklopol> okay umm i tried to read about [ and ] on the wiki but it doesn't make any sense, maybe we could leave those out because i doubt anyone has the time to implement them are they important??
17:27:45 -!- oklopol has quit.
17:27:58 -!- relet has joined.
17:28:06 -!- oklopol has joined.
17:31:00 <fizzie> He is just being silly.
17:31:10 <fizzie> It's an oko thing to do.
17:36:03 <oklopol> i seriously hope that isn't hard to tell :P
17:36:57 <oklopol> although i do seriously want to put the fart language on the wiki, but don't worry i won't
17:36:57 -!- tombom has joined.
17:36:59 <Sgeo> As a programming challenge, [I won the previous one, so I got to make the challenge], I asked the class to implement BF without []
17:37:10 <Sgeo> That was too hard, so I asked just for HQ9+
17:37:33 <oklopol> also this is not a relapse, i started making food.
17:37:33 <Sgeo> [Admittedly, I did make a mistake in what was supposed to be a hint]
17:38:04 <oklopol> or you mean no one cared to do it
17:38:30 <oklopol> or no one was able to write a loop that loops over a string
17:38:34 <Sgeo> I think there's at least one person who would have cared to do it, but he didn't know where to start
17:38:41 <Sgeo> Also: This was in bash, so...
17:39:11 <oklopol> or wait you probably don't have anything else anymore given your age
17:40:06 <oklopol> well anything as fun as a competition doesn't exactly sound like uni stuff
17:40:11 <oklopol> or maybe our uni is boring
17:40:32 <fizzie> oklopol: We have that AI competition, and it's at a university.
17:40:47 <fizzie> There aren't very many courses with that sort of stuff, though.
17:40:57 <Sgeo> It was a small inclass competition
17:41:23 <oklopol> we were supposed to have a competition on one course but then the lecturer never got to writing the platform thing so we ended up doing something retarded
17:41:27 <fizzie> The embedded-systems course have this programmable car, and they have a car-racing competition along a (marked with masking tape) track at the lobby of the CS building.
17:41:46 <fizzie> I didn't do the course, but I've read the instructions once.
17:42:14 -!- cpressey has changed nick to NewNumber2.
17:42:18 <fizzie> There's some sort of an exam before they let you run your stuff on the car, and I think the acceleration values were limited somehow, they're a bit worried about breaking their hardware.
17:42:29 <oklopol> NewNumber2: did coppro retire?
17:43:17 <oklopol> okay now this is starting to seem like a relapse
17:43:47 <fizzie> oklopol: Looks like this: http://www.youtube.com/watch?v=Q9rAEyYqfjw -- except you already went.
17:43:51 <NewNumber2> oklopol: Not unless the subtext of "The Prisoner" is much... deeper... than I had previously considered.
17:44:17 <fizzie> (Caution: not a very exciting video.)
17:47:17 -!- cal153 has quit (Ping timeout: 245 seconds).
17:50:11 <oklopol> NewNumber2: i have no idea what that is, i'm going with coppro
17:51:14 -!- NewNumber2 has changed nick to cpressey.
17:51:52 <oklopol> fizzie: tbh the actual race does not look very exciting :D
17:52:43 <fizzie> oklopol: Yes, it's mostly fun to watch when the cars go all wrong.
17:53:03 <fizzie> Like at the start of http://www.youtube.com/watch#!v=6T_hvq2ptuQ
17:53:19 <fizzie> Still, it's probably more fun to code than to spectate.
17:53:39 <fizzie> (Adding some rocket fuel could make it pretty exciting too.)
17:56:38 <CakeProphet> there should be what is essentially BattleBots but with a live shell as the control.
17:57:27 <CakeProphet> It was this show that aired on TechTV back in the day. US TV station, but I think the show is international (or at least non-US. There were lots of people with English accents)
17:57:39 <CakeProphet> I think BattleBots is like a children's toy or something. :P
17:59:38 -!- kar8nga has joined.
18:00:28 <fizzie> There is a BattleBots show too.
18:01:27 <fizzie> " Wars was a USA based robot competition[1] from 1994-1997. Its considerably modified British equivalent was broadcast on BBC Two from 1998 until 2002, with its final series broadcast on Five in 2003." + "BattleBots is an American company that hosts robot competitions. BattleBots is also the name of the television show created from the competition footage."
18:01:57 <fizzie> There was a Finnish localization of one of those, don't know which one.
18:02:19 <fizzie> Never watched it; using remote controls feels somehow like cheating to me, the bots ought to be autonomous.
18:02:43 -!- atrapado has quit (Quit: Abandonando).
18:03:08 <CakeProphet> but that's difficult in the real-world, outside of your neatly organized memory model. :)
18:04:15 -!- coppro has joined.
18:04:20 <fizzie> Also the robotic overlord issue.
18:05:37 <CakeProphet> When I'm programming sentient AI for robots with self-replication functionality... I never hardcode a manual shutdown mechanism out of sheer principle.
18:06:14 <fizzie> It's your one ticket to fame: you'll be remembered in history books as the betrayer of the human race.
18:06:17 <cpressey> Yes. It would be anti-evolution, which would be anti-life. You're not anti-life, are you? Of course now.
18:06:45 <AnMaster> fizzie, mhm, nice that the gift card works somewhere at least
18:07:13 <fizzie> Somehow I'm reminded of http://isometric.sixsided.org/strips/grease_the_wheel/ for no particular reason.
18:07:51 <AnMaster> okay this is a strange new issue. My thinkpad on resume keeps having backlight set to 0
18:08:19 <AnMaster> very strange, it is on AC, it wasn't behaving like this yesterday
18:09:05 <CakeProphet> Reading college-level students saying the simplistic things I've ever heard kind of causes me to lose hope in humanity.
18:09:15 <AnMaster> another thing, why the fuck does the battery indicator in gnome think the battery is charging. It is not. It is a 50% charge and it won't begin to charge again until it reaches 20%
18:09:28 * AnMaster loves this thinkpad feature. Helps prolong battery life
18:09:59 <AnMaster> echo 100 > /sys/devices/platform/smapi/BAT0/stop_charge_thresh
18:09:59 <AnMaster> echo 20 > /sys/devices/platform/smapi/BAT0/start_charge_thresh
18:12:32 <Phantom_Hoover> AnMaster, stop taunting me with your superior computer.
18:12:34 <coppro> though you can probably do better by not going to 100
18:12:45 <coppro> going to 100 is actually pretty stressful on Li-ion
18:14:16 <AnMaster> coppro, how does sysctl.conf work for /sys?
18:14:34 <AnMaster> coppro, huh, never heard of that
18:14:57 <AnMaster> coppro, where would the file be?
18:14:59 <coppro> you need to install something for it
18:15:09 <AnMaster> coppro, ah, rc.local seems better then
18:16:10 <coppro> in http://www.thinkwiki.org/wiki/Maintenance#Battery_treatment
18:17:05 <CakeProphet> hmmm... I wonder if sentient machines could formulate an ideal government/society.
18:17:38 <CakeProphet> or if they are handicapped in this goal by sentience in the same way humans are.
18:20:10 <cpressey> crappy computers > superior computers
18:20:18 <cpressey> choke on THAT, robot overlords!
18:20:26 -!- cal153 has joined.
18:25:46 -!- augur has quit (Remote host closed the connection).
18:26:01 -!- augur has joined.
18:27:30 <CakeProphet> humans: talking monkeys with talking electric boxes
18:32:10 -!- MigoMipo_ has changed nick to MigoMipo.
18:32:13 -!- MigoMipo has quit (Remote host closed the connection).
18:38:01 -!- MigoMipo has joined.
18:41:19 -!- casdoro1 has joined.
18:46:08 -!- casdoro1 has left (?).
19:04:02 <AnMaster> coppro, btw on my thinkpad the smapi settings seems to stay as long as the battery is in, while that page on thinkwiki suggests it only stays as long as AC is in
19:24:10 -!- kar8nga has quit (Remote host closed the connection).
19:26:57 -!- kar8nga has joined.
19:36:07 -!- kar8nga has quit (Remote host closed the connection).
19:37:54 -!- kar8nga has joined.
19:38:38 -!- Geekthras has quit (Ping timeout: 276 seconds).
19:41:59 -!- kar8nga has quit (Remote host closed the connection).
19:44:56 -!- Geekthras has joined.
19:49:02 <Sgeo> Stony Brook doesn't have a Computer Science MS, afaict
19:51:55 <AnMaster> isn't MS something you see on boats?
19:52:44 <AnMaster> ah yes, "motor ship" according to wikipedia
19:52:53 <AnMaster> a compsci boat sounds fun though
19:53:13 <AnMaster> but probably not what Sgeo meant
19:53:31 * Sgeo meant master's
19:54:30 <AnMaster> Sgeo, change university then? I mean, once I'm done with my bachelor, I'm not going to continue at this university. Only compsci master they have is about AI, and that is boring
19:54:39 <AnMaster> so going to switch university then.
19:54:58 <coppro> hey, I know several AI Ph.D.s or candidates
19:55:25 <AnMaster> coppro, yes but I find the subject boring
19:55:46 <Sgeo> I think it's just that I want some sort of CompSci degree, and not "Computer Programming/Information Systems"
19:56:17 <cpressey> Sgeo: That's all the BSc covers these days at most schools, isn't it.
19:57:21 <cpressey> Look, here's Java! And this is what an algorithm looks like! And an OS! Oh, now spend 60 hours straight on this "software engineering" project! Great, now go away!
19:57:43 -!- sebbu2 has joined.
19:59:16 -!- sebbu has quit (Ping timeout: 265 seconds).
19:59:17 -!- sebbu2 has changed nick to sebbu.
20:05:17 <Phantom_Hoover> cpressey, this is why I have firmly resolved not to go into CS.
20:06:32 <cpressey> Phantom_Hoover: Good move. Although what other fields are there for a young, aspiring computer geek? EE will rot your brain, while Mathematics will destroy your soul...
20:07:41 <cpressey> Always good to be prepared for these things.
20:07:43 <AnMaster> Phantom_Hoover, not that either
20:08:16 <AnMaster> cpressey, there was some basic EE stuff in this CS course. I can easily see how more of it would not be a good idea
20:08:34 <CakeProphet> but I think I'll probably stick to Computer Science, as I know for certain that I enjoy it.
20:08:37 <AnMaster> mathematics yeah... I hate analysis
20:09:14 <CakeProphet> though computer engineers have a higher median salary... CS salary isn't really that shabby either.
20:09:54 <AnMaster> CakeProphet, this compsci is closer to compeng to tell the truth. But most are nowdays
20:10:21 <AnMaster> extrapolating from compsci, computer engineering must be compeng
20:10:24 <CakeProphet> AnMaster: really? My university's programming is pretty much strictly programming.
20:10:42 <AnMaster> CakeProphet, yes but that isn't compsci really. compsci is way more theoretical
20:10:56 <coppro> Proper compsci is a math discipline
20:11:00 <CakeProphet> AnMaster: ah. Then perhaps you're talking about Software Engineering
20:11:07 <CakeProphet> Computer Engineering is basically EE but specialized.
20:11:24 <coppro> You guys want software engineering
20:11:40 <CakeProphet> at my university there's practicalyly no difference between software engineering and compsci
20:11:54 <AnMaster> coppro, I want compsci but I get a tiny bit of compsci with mostly softeng mixed up with some compeng :/
20:11:59 <CakeProphet> but I mean... it would be ridiculous to dump someone in a theoretical computer science class without teaching them programming basics first.
20:12:02 <cpressey> coppro: The observation was that "proper compsci" is really rare these days, at most schools
20:12:25 <coppro> thankfully I'm going to UW
20:12:42 <CakeProphet> ...I told you. We need an acronym counting bot.
20:12:47 <coppro> University of Waterloo
20:13:29 <CakeProphet> I do believe later courses will be more thereotical. I have to take a complexity theory class... though I'm already way ahead of that.
20:14:16 <Sgeo> Is it sad that my best laugh in a long time was in response to someone posting a picture of a clam to someone saying "I'm sure a this thread will contain a clam and reasonable discussion of religion and Apple."
20:14:38 <AnMaster> btw, according to powertop, ubuntu lucid is by default way worse than jaunty
20:14:50 <AnMaster> and not sure how to fix most of that
20:14:57 <CakeProphet> but I don't really go to university for the knowledge (though it is nice to have knowledgable professors to discuss things with). I'm purely interested in the credentials. I can learn at my own leisure just about whatever I want.
20:15:40 <CakeProphet> so I'm not too upset that my compsci degree won't include much theoretics.
20:15:50 <AnMaster> Sgeo, don't know, however: it is sad that you ask us about opinions on such things all the time
20:16:02 <cpressey> CakeProphet: I went for the piece of paper too. But made good use of the library access while I was there. Plus *some* of the courses were not lame.
20:16:08 <CakeProphet> hmmm, does anyone here actually have a BA or higher in compsci?
20:16:21 <CakeProphet> cpressey: the library access is quite nice. And yes, I agree.
20:16:46 <cpressey> It's possible. Some Uni's treat it and math as arts.
20:16:49 <coppro> I'm going for a BMath myself
20:16:57 <cpressey> Well, I mean I can see math not being a science.
20:16:59 <AnMaster> it isn't even called that in Sweden
20:17:45 <CakeProphet> cpressey: Computer Music Journal is great. :D
20:18:02 <CakeProphet> gives me ideas for all the awesome experimental computer music I've yet to make.
20:18:10 <CakeProphet> AnMaster: Talking about awesome libraries.
20:18:22 <AnMaster> CakeProphet, I meant that journal thingy
20:19:08 <CakeProphet> I don't think the newer ones are on there though. Academia likes to be non-open for whatever reason.
20:20:00 <CakeProphet> One project I have in mind is transforming a wii-mote into a musical instruments / MIDI controller
20:20:33 <CakeProphet> ...do you have your browser set to ask you about such things?
20:20:42 <AnMaster> yes, and I click no all the time
20:21:07 <CakeProphet> I've yet to witness any detrimental impact to accepting all cookies yet.
20:22:03 <CakeProphet> AnMaster: ah, other way around actually. The online version has new issues but not old.
20:22:35 <CakeProphet> but they're all great. All interesting ways to create and process digital audio signals.
20:23:11 <CakeProphet> I really wish there was a somewhat decent DSP language/library out there. Haskell has one but I'm not sure if it's really up-to-par with domain-specific languages like csound.
20:23:40 <CakeProphet> but csound has crappy general-purpose support in terms of state and control structures. Just really solid digital signal processing.
20:28:31 <CakeProphet> one day I'd like to have a really nice composition environment. Like a shell of some kind connected to a musical score with optional MIDI output.
20:29:24 <cpressey> CakeProphet: Although I used to compose, I've never been too heavily into "computer music", such as it is.
20:30:05 <cpressey> It's just that I never learned piano, so a computer's the only instrument I can compose on :)
20:30:07 <AnMaster> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/524281 <-- argh
20:30:23 <AnMaster> this will be a PITA when using it on battery for a while
20:31:23 <CakeProphet> cpressey: I'm not too keen on most of it. I simply use it as educational material. I vastly prefer more popular varieties of electronic music (well, that's not entirely true either. Only the "good" stuff, of course. :P )
20:31:56 -!- EgoBot has quit (Ping timeout: 276 seconds).
20:32:18 <CakeProphet> and of course, I don't solely like electronic music.
20:32:35 -!- HackEgo has quit (Ping timeout: 276 seconds).
20:34:45 <cpressey> CakeProphet: I think all music should ship with a parallel MIDI version, so that we can write visualizers that aren't lame.
20:35:09 <CakeProphet> ha... yeah that'd be pretty sweet. At least for electronic music where MIDI is used and makes sense
20:35:21 -!- oerjan has joined.
20:35:25 <CakeProphet> doesn't work too well with analog sources. Still doing the same pattern matching you'd be doing in visualizer software.
20:35:32 <cpressey> Well, even for non-electronic, it would at least give it a "click track" it could follow.
20:35:38 <CakeProphet> there are some really impressive visualizers these days though. Have yo seen Milkdrop?
20:35:47 <oerjan> Phantom_Hoover: ok i give up
20:36:09 <cpressey> CakeProphet: No, I have to say I've not been following development in that area.
20:36:18 * oerjan kneels and kisses Phantom_Hoover's disembodied toes
20:36:44 <CakeProphet> cpressey: it's pretty stunning. Very, uh, trippy. It syncs to tempo automatically. Don't even ask me how.
20:37:17 * oerjan shoots Phantom_Hoover with a bow and arrow, which passes straight through and hits CakeProphet
20:37:58 <CakeProphet> it is dark. you are likely to be eaten by a grue.
20:38:49 <CakeProphet> Is there an IRC client that just lets you pipe to/from stdio?
20:38:55 <CakeProphet> would make writing an IRC bot far more convenient.
20:39:46 <oerjan> Phantom_Hoover: aka grue@something
20:40:00 * oerjan cannot find him in any recent logs
20:40:50 <CakeProphet> coppro: well, sort of. nc would work but then I might as well just do it via language.
20:41:55 <coppro> there's not much else to abstract
20:42:00 <coppro> IRC protocol is very simple
20:42:21 <cpressey> CakeProphet: There is apparently a piece of spyware out there called "IRC Pipe"...
20:42:38 <cpressey> Not suggesting you build on top of spyware tho.
20:45:04 <oerjan> <oklopol> so what's the puzzle today, oerjan
20:45:13 <CakeProphet> cpressey: psh, what are you talking about. That's fool-proof
20:46:49 <oerjan> oklopol: i found this nice math, computing and complexity blog btw, http://rjlipton.wordpress.com/
20:47:12 <oerjan> i learned about a result there that is _almost_ like a puzzle
20:47:35 * oerjan notes oklopol is long time idle
20:50:31 <oerjan> basically, what boolean functions can you compute by chaining computations that can only look at one bit of the input each, and only pass a _bounded_ amount of data to the next one, but unlike an FSA you can have several of them looking at the same input bit
20:50:39 <oerjan> http://rjlipton.wordpress.com/2009/03/02/barrington-gets-simple/
20:51:36 <CakeProphet> Erik Zeeman conjectured that one cannot untie a knot on a four-sphere. I am not a topologist, but it seems like a reasonable conjecture.
20:53:07 <Phantom_Hoover> You can't untie knots in 3-space, and the "surface" of a 4-sphere is 3-space.
20:54:18 <CakeProphet> I have no clue what a 4-sphere is. Unless it's a 4-dimensional sphere. And then I still have trouble imagining it... but will simply recognize that it could be defined.
20:54:43 <CakeProphet> The Truth: After trying to prove this for almost ten years, one day he worked on the opposite direction, and solved it in hours.
20:54:45 <oerjan> and surprisingly the answer is "all of them, using only 5 possible passing-on values" (the 5 i deduced myself from the article)
20:55:16 <Phantom_Hoover> It's impossible to directly visualise, but you can model it mentally pretty easily.
20:55:53 <oerjan> CakeProphet: iirc from the comments that knot conjecture was wrongly quoted, it should be a kind of 2-dimension knot or something
20:56:03 <fizzie> There's also nsound, which is csound-inspired but tries to make the programming parts better. Don't have any personal experiences with it. (There's this one audio-processing-tuned library I distinctly remember, but I've completely forgotten the name, trying to Google for it.)
20:56:35 <oerjan> well some 2D generalization of a knot
20:57:17 <fizzie> http://en.wikipedia.org/wiki/Comparison_of_audio_synthesis_environments has some sort of list. (But I can't find the one I ran across.)
20:57:29 <cpressey> fizzie, CakeProphet: something crossed my mind recently re computer music. Either of you know of any musical instrument simulators? i.e. they take the laws of physics, and a description of the instrument, and produce the noise it makes?
20:57:29 <CakeProphet> fizzie: oh really? Haven't heard of it. I'll look into it. Sounds like what I want (though I could probably just use some IPC, or attempt to learn the Haskell library)
20:57:42 <oerjan> Phantom_Hoover: i _think_ that a 4-sphere may actually be a sphere in R^5, so it's shifted by 1.
20:58:03 <Phantom_Hoover> CakeProphet, fizzie, cpressey, We're mathematicsing now.
20:58:34 <fizzie> cpressey: The acoustics lab at our university has been building some string instrument simulators, but of course it's university research, so there's no real products, just papers.
20:58:44 <AnMaster> oerjan, what about a one-sphere?
20:59:05 <cpressey> fizzie: I was wondering if it would be particularly hard to simulate an electric guitar...
20:59:05 <oerjan> AnMaster: that would be a circle, then
20:59:15 <AnMaster> oerjan, how do you get a dot then? 0-sphere?
20:59:20 <Phantom_Hoover> oerjan, on the (4D) surface of a 4-sphere you can always untie a knot.
20:59:35 <cpressey> It's just Maxwell's equations and a vibrating conductor and a stationary magnet, really.
20:59:40 <oerjan> "For any natural number n, an n-sphere of radius r is defined as the set of points in (n + 1)-dimensional Euclidean space which are at distance r from a central point, where the radius r may be any positive real number."
21:00:11 <AnMaster> oerjan, but that causes a headache for 1-dimensional!
21:00:19 <oerjan> AnMaster: a 0-sphere is 2 dots
21:00:40 <AnMaster> oerjan, right. what about a (-1)-sphere?
21:01:12 <AnMaster> CakeProphet, since infinity + 1 = infinity it will actually match in that special case ;P
21:01:22 <oerjan> a (-1)-sphere is the empty set. and iirc there were some theorems in algebraic topology that fit with all of this.
21:01:35 <fizzie> cpressey: For the acoustic guitar, the corresponding our-university project page is http://www.acoustics.hut.fi/research/asp/aguitar/ -- no publications past 2005, so perhaps it's not very active right now.
21:01:47 <AnMaster> cpressey, aleph-zero is a more common name iirc
21:02:10 <CakeProphet> fizzie: Reaktor is a very nice environment for DSP. Unfortunately it's $579.
21:02:28 <AnMaster> Phantom_Hoover, aleph zero is more common than that too
21:02:28 <Phantom_Hoover> Although "nought" is apparently an archaism in the US, so.
21:02:43 <AnMaster> Phantom_Hoover, it isn't in the UK iirc?
21:02:49 <AnMaster> and that is all that matters IMO
21:02:53 <fizzie> I'm sure someone's tried to model the electric sli^H^H^guitar too.
21:03:10 <Phantom_Hoover> AnMaster, but there are American mathematicians, much as I hate to admit it.
21:03:13 <cpressey> fizzie: Interesting. I was thinking electric would be easier, because there's no air or sounding chamber involved -- the electric waveform gets translated directly to the sound wave.
21:03:54 <CakeProphet> Phantom_Hoover: the only time I hear "nought" is in the phrase "all for nought"
21:03:56 <AnMaster> Phantom_Hoover, sure. aleph-zero is neutral and fine IMO
21:04:21 <AnMaster> Phantom_Hoover, what does? aleph-zero?
21:04:59 <fizzie> cpressey: Oh, and the parent page -- http://www.acoustics.hut.fi/research/asp/ -- has links to kantele synthesis; can't say I'd be very surprised if it turned out no-one else than Finns were doing *that*.
21:05:06 <oerjan> Phantom_Hoover: a 1d (ordinary) knot is an embedding of a circle into R^3. since a circle is a 1-sphere, a 2d knot would probably be an embedding of a 2-sphere (i.e. ordinary sphere) into R^4.
21:05:45 <CakeProphet> cpressey: well, it does get transferred directly. But you still need to synthesize the resonance frequencies and such. The somewhat naive way to do it is to use a waveshaping function (aka distortion) on an FM synthesized wave.
21:06:12 <oerjan> Phantom_Hoover: oh. well that would be 2 points then. not very interesting :D
21:06:18 <fizzie> AnMaster: "The kantele is a traditional Finnish plucked string instrument with five metal strings in its basic form."
21:06:19 <CakeProphet> cpressey: but you could get more complicated. I don't know much about physical modelling techniques unfortunately.
21:06:43 <cpressey> CakeProphet: I guess you're right - physically modelling a vibrating string is still, uh, non-trivial.
21:06:51 <Phantom_Hoover> And the thing about untying knots on a 4-sphere is therefore presumably referring to 2-knots.
21:06:51 <AnMaster> btw, I quite like the norwegian hardangerfela (spelling?)
21:07:13 <fizzie> And: "Research and synthesis of the tanbur, a traditional Turkish long-necked lute, the ud, a short-necked arabic lute, and the Renaissance lute". Our tax money at work there.
21:07:29 <Phantom_Hoover> And any knot that is untyable in flat 4-space is untyable on a 4-sphere.
21:08:30 <AnMaster> Phantom_Hoover, what does untying a knot mean? getting a straight piece of string? or getting a single loop?
21:09:00 <oerjan> AnMaster: hardingfele is the usual word in norwegian
21:09:13 <oerjan> but i see danish wp uses hardanger
21:09:34 <AnMaster> oerjan, and Swedish one? Don't have browser handy atm
21:09:35 <oerjan> AnMaster: it's from the same geography name, anyhow
21:09:36 <CakeProphet> SuperCollider looks interesting for sound synthesis. Has a decent programming language too.
21:09:48 <AnMaster> oerjan, yeah anyway, quite a nice sound.
21:09:54 <AnMaster> fizzie, you should get someone to synth that
21:10:22 <oerjan> AnMaster: ah the swedish actually explains the name development
21:10:26 <AnMaster> CakeProphet, just watch out for black holes
21:10:38 <oerjan> "Hardingfelan (av den numera formella namnformen Hardangerfela, efter det norska landskapet Hardanger; i dagligt tal vanligen endast Harding)"
21:10:47 <fizzie> I could suggest the acoustics lab people; that second set of strings sounds interesting.
21:11:04 <AnMaster> oerjan, shame on you for that not being on the norwegian wiki
21:11:16 <AnMaster> oerjan, hm does the norwegian wp use bokmål or nynorsk?
21:12:12 <AnMaster> fizzie, btw those synthed instruments, there is no software to hook it up to a midi program is there?
21:12:54 <fizzie> AnMaster: I don't think they've managed to release much code. But you can always reimplement the algorithms. :p
21:13:12 <AnMaster> fizzie, true. Don't have access to all those journals though
21:13:51 <cpressey> CakeProphet: Do you (or anyone else here) do any (non-algorithmic) composition? If so, I'd be interested in what software you use.
21:13:52 <fizzie> AnMaster: A lot of their research is available from acoustics.hut.fi; conference papers, anyway. Journal articles might not be.
21:14:08 <cpressey> Or if anyone remembers what software Gregor uses :)
21:14:24 <oerjan> Phantom_Hoover: http://en.wikipedia.org/wiki/Knot_theory#Higher_dimensions
21:14:59 <cpressey> Assuming he's not here right now. Happy to be proven wrong of course.
21:15:05 <oerjan> AnMaster: there are two distinct wikipedias for bokmål and nynorsk
21:15:16 <AnMaster> cpressey, http://codu.org/music/
21:15:16 <fizzie> AnMaster: Since conference papers are usually 4 pages and you need to present results and stuff, not just methods, there's usually a *very* compact (and often incomplete) description of the algorithms involved, though.
21:16:01 <oerjan> AnMaster: in fact the nynorsk one has some articles with variants in "høgnorsk", a more archaic form
21:16:17 <CakeProphet> cpressey: not much composition at all, really. But I use Ableton Live for general-purpose recording of electronic music. Though it's not technically for "composition" (no musical score or anything). It's an expensive program (or rather, an expensive copyright. torrents ftw :P)
21:16:45 * CakeProphet has broken several laws in his lifetime. Frequently.
21:16:53 <oerjan> AnMaster: theoretically the bokmål one should have corresponding "riksmål" variants but i cannot recall seeing that
21:17:04 <fizzie> We all have recommendations to get also source code published whenever possible, but due to economical realities (papers == funding) it doesn't happen so often.
21:17:09 <oerjan> we take our language history _seriously_ :D
21:17:10 <AnMaster> oerjan, good the English aren't like you. Or we would not only have en-us.wikipedia.org en-gb.wikipedia.org but also en-cockney.wikipedia.org and so on
21:17:22 <oerjan> AnMaster: there is simple english, though
21:17:41 <fizzie> http://www.acoustics.hut.fi/demos/ has guitar synthesis audio clips, however. :p
21:17:42 <AnMaster> oerjan, true, but that is for a somewhat different purpose and reason
21:18:29 <cpressey> I used to use an old version of Cakewalk. One that writes files that no subsequent version of Cakewalk can read...
21:18:46 <fizzie> "The song was performed live with a MIDI keyboard -- "; so obviously they *have* wired the code into a MIDI pipe, they just don't let anyone else have it. :p
21:18:46 <AnMaster> fizzie, but doesn't sound sampled, not high enough quality for that
21:19:09 <cpressey> And... codu.org isn't responding for me.
21:19:38 <CakeProphet> the difficult of simulation of acoustic instruments isn't really the spectrum of the waveform or even the mathematical complexity of constructing said waveform. It's the dynamics we hear from the human performer. You can get a lot of expressiveness out of a good hardware controller. Barring that, a sophisticated attack-sustain-decay-release for each overtone does the job.
21:19:44 <AnMaster> cpressey, you should have been quicker!
21:20:18 * Phantom_Hoover realises that there are now websites dedicated to giving Facebook users endless mountains of crap to "like".
21:20:26 <AnMaster> fizzie, a bit sad that sampled is still better than their stuff
21:21:10 <fizzie> Sampled is reality; it's hard to beat that.
21:21:28 <cpressey> And FluidSynth for the mp3 rendering?
21:21:29 <AnMaster> cpressey, some are recorded live iirc?
21:21:37 <AnMaster> cpressey, if you mean codu.org
21:21:44 <oerjan> Phantom_Hoover: http://rjlipton.wordpress.com/2010/06/19/guessing-the-truth/#comment-3706
21:21:48 <AnMaster> cpressey, and he uses good samples for the fluidsynth cases
21:21:51 <cpressey> I mean "what software does Gregor use"
21:21:52 <CakeProphet> fizzie: if you sample guitar tones and then play them back with MIDI keyboard... it will not sound like a guitar.
21:22:17 <CakeProphet> it will sound like guitar tones mapped to keyboard keys.
21:22:39 <AnMaster> cpressey, rosegarden yes iirc. And sometimes he uses fluidsynth with good samples, in other cases he records himself playing
21:22:41 <Phantom_Hoover> oerjan, so it's not as clear cut with 2-knots in 5-space?
21:23:14 <cpressey> CakeProphet: But if you can get a human performer behind the controls of a synthesized instrument...
21:23:19 <CakeProphet> cpressey: AnMaster and probably JACK, of course.
21:24:11 <fizzie> Current state-of-the-art (if you just need a single voice and don't mind horrible amounts of work to prepare it) in speech synthesis (AFAIK) is still concatenative synthesis, basically sewing small snippets of recorded speech together (with quite a few tricks in the concatenation and segment-selection, but still).
21:24:15 <CakeProphet> Jack is basically a plug-in interface. You take programs that support JACK and plug together their inputs/outputs however you like.
21:24:18 <cpressey> CakeProphet: But JACK isn't really a... tool, is it? I thought it was more like ... glue.
21:24:38 <AnMaster> CakeProphet, I know what jack is. I used it. It is however a bitch to get working
21:26:37 <cpressey> I'm thinking of getting back into composition. I don't have a KB, but my composition paradigm doesn't really make full use of one anyway -- I grew up on things like SidPic and Deluxe Music Construction Set and MED (an Amiga tracker editor). I'm hoping RoseGarden will fit into this paradigm.
21:27:06 <CakeProphet> cpressey: buy a wii-mote... or make something similar. get an IR port of some kind. epicness ensues.
21:27:19 <CakeProphet> I saw a youtube video that used a wii-mote to implement a virtual drum machine.
21:27:24 <cpressey> CakeProphet: *Composition*, not *conducting*. Duh.
21:27:32 <CakeProphet> but there are other uses. Continuous pitch ranges and such
21:27:42 <CakeProphet> cpressey: oh, right. psh. performance is so much more fun though.
21:27:53 <oerjan> Phantom_Hoover: well i don't actually know this stuff myself, just linking :D
21:28:27 <cpressey> Well... I did, until recently, have a physical trombone... virtually replace it with a wii-bone? Hmm...
21:29:41 <Phantom_Hoover> cpressey, I once got FlightGear running with a Wii remote as the joystick.
21:30:12 <AnMaster> CakeProphet, I never got the jokes about wii and erotic... *shrug*
21:30:25 <cpressey> Phantom_Hoover: Sounds like a fun new way to crash into mountains!
21:30:48 <Phantom_Hoover> Mainly because I tested it in KSFO, where there aren't any mountains.
21:31:01 <AnMaster> I would use my nice saitek x52 pro for flying
21:31:28 <AnMaster> I don't crash into mountains :P
21:31:29 <CakeProphet> depending on whether or not you want to record electronic style music... have some kind of control surface/device is very very handy. In Ableton Live you can assign MIDI knobs/sliders to various virtual parameters. I'm not sure if you could use this approach with Ableton Live, but you could make a software MIDI wrapper over the IR input and then plug that into JACK or something.
21:31:57 <fizzie> The PS3 sixaxis reports itself as a HID joystick with 28 analog axes, according to what I read during the recent USB discussion. That should be enough for everyone! (If they were usable for anything; it's just that most of the buttons on the thing are pressure-sensitive, and therefore reported both as on/off buttons as well as analog axes.)
21:31:58 <CakeProphet> and then, bam, MIDI wii controller/instrument.
21:32:35 <AnMaster> fizzie, while the saitek x52 pro only has 10 axis, it has 38 buttons
21:32:42 <CakeProphet> I could awkwardly control so many musical paramters with my thumb... I can't even imagine.
21:32:45 <AnMaster> well some are on the same, but the button pressed different ways
21:33:21 <AnMaster> Phantom_Hoover, my desktop is crap as I said, sempron 3300+ with geforce 7600
21:33:34 <AnMaster> Phantom_Hoover, however it does have a SB Live 5.1 sound card
21:33:44 <fizzie> AnMaster: 8 pressure-sensitive buttons and a pressure-sensitive d-pad (so 4 more). And then two analog sticks (two more axes of freedom), 2 analog trigger buttons, three digital on/off-only buttons, and 6 axes of motion-sensing (I don't exactly know what those are).
21:33:59 <CakeProphet> AnMaster: I bet my friend has a more awesome card. M-audio 1010LT (I think)
21:34:08 <cpressey> CakeProphet: Actually... I *am* interested in *some* algorithmic aspects -- I would probably go to code before I would go to recording analog inputs to do pitch bends and so forth. But I'm not sure. I've been fooling around with chiptune-like effects in C64 6502 recently. Not in seriousness.
21:34:12 <AnMaster> CakeProphet, well, most awesome consumer level card maybe
21:34:21 <AnMaster> CakeProphet, m-audio sounds like a pro card
21:34:25 <cpressey> I dunno, maybe for phase sweep.
21:34:52 <CakeProphet> AnMaster: It has like 10 analog inputs, a mic-input with pre-amp, MIDI inputs, and I think 4 digital inputs. It's a mass of wires coming out of the back of his desktop case.
21:35:11 <AnMaster> CakeProphet, hardware mixer? hardware midi?
21:35:25 <AnMaster> CakeProphet, and hardware midi I mean synth in hardware
21:35:31 <AnMaster> that you can load samples into
21:35:55 <CakeProphet> dunno about either of thus. Dunno why you would want a hardware synth on your soundcard. :P
21:36:10 <cpressey> Phantom_Hoover: Topological music.
21:36:13 <CakeProphet> at least there's a software interface for a mixer... dunno how it's implemented though.
21:36:16 <AnMaster> CakeProphet, I would be surprised if it didn't mix in hardware
21:36:48 <AnMaster> CakeProphet, after all, hardware midi means it offloads the CPU, which is nice when you have a 72 MB soundfont
21:36:49 <cpressey> We don't describe notes in terms of their concrete relationships like pitch and duration, we just... er...
21:37:04 <fizzie> 1010LT: http://www.m-audio.com/images/global/product_pics/big/delta1010lt_v2008.jpg
21:37:07 <fizzie> Nice connector-squid there.
21:37:09 <CakeProphet> AnMaster: but yeah, it's the sickest sound card I've seen so far. When I get my future desktop, and can spare the cash, I'll probably get one. It's basically the best sound card for professional recording I've found.
21:37:10 * cpressey tries to figure out what it would mean for a set of music events to be "connected"
21:38:00 <AnMaster> the word for it slipped my mind atm
21:38:11 <CakeProphet> generally if they similar attack/decay times we perceive a musical event as atomic.
21:39:21 <cpressey> So silence between notes = holes in a deformable object, maybe.
21:39:28 <fizzie> AnMaster: According to the product description I saw when looking for the photo, "All Delta cards contain a 36-bit embedded DSP enabling a software-driven patchbay/router for all analog and digital I/O—all with extremely fast throughput for low-latency software monitoring." In other words, yes, it has a "hardware mixer" (and more), as long as you count the completely programmable DSP on the sound card as part of hardware.
21:39:37 <CakeProphet> a chord is one musical event. Though with chords we can pick out the individual notes. There are actually distances in pitch called "critical bands" that we perceive as seperate tones. Anything inside a critical band sounds like a single tone.
21:39:48 <AnMaster> cpressey, sounds like stretching concepts way past the point where it broke
21:40:03 <cpressey> AnMaster: What channel do you think this is, man?
21:40:09 <fizzie> Phantom_Hoover: Probably a few spare bits to avoid clippingsy problems.
21:40:32 <oerjan> Phantom_Hoover: however, the fact that wikipedia section said the results for smooth knots were different from piecewise linear ones makes me suspect the higher-dimensional cases are _very_ subtle
21:41:34 <CakeProphet> Another propery of critical bands is they become wider in the lower octaves. That is why chords in lower octaves need to be spaced out in order to sound like chords... otherwise the spectra of each note will blend into one pitch and we will perceive it as dissonant .
21:42:15 <Phantom_Hoover> cpressey, CakeProphet, stop having such similar names.
21:42:21 <AnMaster> CakeProphet, you just made me think "three man critical band" ;P
21:42:52 <fizzie> I do vaguely recall that there was a lot that *could* be done by reprogramming the EMU10k DSP chip on SB Lives, it just wasn't very documented stuff.
21:43:04 <CakeProphet> has something to do with actual physical space on the basilar membrane of the ear, and which sections are used to detect given parts of our psychoacoustic spectrum.
21:43:09 <coppro> I wonder what the odds are that if you pick some topic completely at random a) someone in this channel will be an expert or b) someone in this channel could get an expert in the channel in 48 hours
21:43:15 <AnMaster> fizzie, but pretty sure that is the case yes
21:43:36 <AnMaster> coppro, I know it isn't 1 at least
21:43:44 <fizzie> E-MU does pro-level audio stuff, after all.
21:43:59 <AnMaster> coppro, there have been things that no one in here knew about. Forgot what though
21:44:07 <coppro> AnMaster: nonetheless, it's pretty high compared to random groupings of people
21:44:25 <coppro> hey, I have one of those
21:44:34 <AnMaster> coppro, also I suspect a lot of us are experts at appearing like experts
21:44:40 <AnMaster> <oerjan> Phantom_Hoover: well i don't actually know this stuff myself, just linking :D
21:44:44 <fizzie> "Real life", does it have something to do with the 8088 real mode?
21:44:48 <coppro> but then again, what is the definition of expert?
21:45:05 <AnMaster> coppro, we need a philosophy expert for that
21:45:06 <coppro> for instance, though I have no formal training, I consider myself an expert when it comes to c++
21:45:11 <AnMaster> anyone? please raise your hand!
21:45:17 <Phantom_Hoover> And then protected life would be 32-bit and a pig to set up.
21:45:53 <AnMaster> coppro, and I consider myself pretty good at C, though ais523 might beat me there.
21:46:07 <coppro> I don't mean coding C++, just knowledge
21:46:17 <AnMaster> coppro, same, hasn't been coding C much recently
21:46:36 <AnMaster> though I guess I'm the resident erlang expert. Since cpressey isn't here all the time
21:46:55 <coppro> man, I'm so fantastically bad at Erlang
21:47:16 <AnMaster> coppro, you know erlang? I didn't know
21:47:47 -!- fizzie has quit (Quit: jumpin' jumpin').
21:47:49 -!- fizzie has joined.
21:47:50 <cpressey> I'm not an expert at anything.
21:48:02 <AnMaster> cpressey, at the languages you created?
21:48:17 <AnMaster> hm wait, Deewiant probably knows more about befunge98 than anyone else here
21:48:17 -!- fungot has joined.
21:48:27 <fungot> AnMaster: stuff like " thisfunction2 thislineofcode")) but have stan actually reference pete.
21:48:31 <fizzie> I consider myself the resident fungot expert, but that's a rather specialized field of knowledge.
21:48:31 <fungot> fizzie: " eeek"? yeah, do is in r5rs?? i was just trying
21:48:42 <AnMaster> fizzie, and I won't challenge that title
21:48:47 <oerjan> Phantom_Hoover: however i've also read in general that there are many ways in which dimensions 3 and 4 are the _hardest_ to analyze, more so than both lower and higher ones.
21:48:51 <CakeProphet> ...I probably have the most experience programming while under various hallucinogens on this channel.. but I'm not really sure if that is a skill or a psychic power.
21:49:14 <coppro> AnMaster: I know it, but I'm really bad at programming it
21:49:36 <coppro> fizzie: A bit esoteric, wouldn't you say?
21:49:42 <AnMaster> coppro, not too bad at that myself I think. Though I might overestimate my skills. Hard to tell.
21:50:06 <AnMaster> coppro, but that is what this channel is about
21:50:31 <AnMaster> coppro, heh I realised why ais is holding back feather. So he can be the world's leading expert on it
21:50:38 <AnMaster> since no one else knows much about it yet
21:51:01 <AnMaster> Phantom_Hoover, time traveling language that lets you modify the syntax of the language in the past
21:51:24 <coppro> I have some sneaking suspicion that I've inadvertently become the world's leading expert on some esoteric portion of C++
21:51:29 <AnMaster> Phantom_Hoover, it does cause not end of headaches it seems
21:51:38 <coppro> if not, definitely second place
21:51:52 <AnMaster> Phantom_Hoover, the spec isn't finished
21:52:00 <oerjan> <AnMaster> anyone? please raise your hand! <-- and if there are any experts here on telekinesis, raise _my_ hand </old joke>
21:52:02 <AnMaster> coppro, who would hold the first place?
21:52:10 <AnMaster> Phantom_Hoover, on irc, that is all
21:52:21 <coppro> AnMaster: A David Vandervoorde.
21:52:29 <CakeProphet> ...I am an expert of the Pokemon gameboy games. But nothing else, not the television series or anything related to it.
21:52:41 <coppro> the only other person I know to have implemented the attributes specification from C++0x
21:52:56 <coppro> and, unlike me, actually a member of the standard committee
21:52:56 <AnMaster> Phantom_Hoover, and well, ais said he needed to implement it while working on the spec. Since he wasn't sure what would work
21:53:08 <AnMaster> Phantom_Hoover, and that hasn't happened yet
21:53:13 <oerjan> Phantom_Hoover: something like higher dimensions having more wiggle room for stuff
21:54:05 <AnMaster> oerjan, of course. There is an extra dimension to wiggle in
21:54:11 <AnMaster> of course there is more wiggle room thus
21:54:47 <AnMaster> Phantom_Hoover, mathematically they could be anything you want
21:54:54 <AnMaster> Phantom_Hoover, plus: tiny > nothing
21:55:06 <CakeProphet> consider dimensions that ate only one-element sets. WHAT THEN?
21:55:15 <oerjan> Phantom_Hoover: intuitively, 3 and 4 are big enough to let you make complicated stuff but not big enough to let you straighten it out. or something.
21:55:23 <AnMaster> Phantom_Hoover, true but better than nothing
21:55:25 <oerjan> (very vague intuition, there :D)
21:55:59 <CakeProphet> I'm trying to imagine where the fourth axis would be in a world of 4D space. :D
21:56:25 <AnMaster> I wonder what a 4D game would be like.
21:56:47 <AnMaster> Phantom_Hoover, yeah but a space game iirc?
21:56:54 <AnMaster> Phantom_Hoover, I meant, something like a RPG
21:56:57 <Phantom_Hoover> CakeProphet, it goes at right angles to everything else.
21:57:25 <Phantom_Hoover> AnMaster, how would a 4th dimension affect the RPG mechanic?
21:57:30 <CakeProphet> Phantom_Hoover: I need more space than the infinite space I imagine things in...
21:57:33 <cpressey> CakeProphet: "where" is a limiting concept, obviously
21:57:41 <AnMaster> Phantom_Hoover, think what theifs could do!
21:57:49 <AnMaster> Phantom_Hoover, get into to a box without picking the lock
21:58:02 * cpressey leans back, loses balance, and falls through the wall
21:58:14 <oerjan> Phantom_Hoover: topologically, the _size_ of a dimension doesn't matter. although whether it curves back on itself might.
21:58:39 <AnMaster> Phantom_Hoover, well, can you biject C onto R? that should result in the same thing, shouldn't it?
21:58:46 <Phantom_Hoover> cpressey, you have to go at right angles to reality to fall though things.
21:58:58 <AnMaster> Phantom_Hoover, and iirc they are bijectable, forgot how though
21:59:05 <cpressey> biject R^2 onto R? I don't *think* so...
21:59:29 <CakeProphet> perhaps P!=NP, and instead P=R^3 ....so really when I write to a hash table I'm really teleporting through space.
21:59:29 <AnMaster> cpressey, both are uncountably infinite, could be tricky to diagonalize?
21:59:31 <oerjan> Phantom_Hoover: you can biject R^2 onto R as long as you don't require continuity
22:00:17 <AnMaster> CakeProphet, even below OGS (Oerjan Groan Standard)
22:00:33 <oerjan> AnMaster: rubik's hypercube might be an interesting game. hm, i should google if it exists already
22:00:44 <CakeProphet> AnMaster: More of a thought experiment to imagine what that would imply.
22:00:48 <AnMaster> oerjan, I never got the hang of the 3D version so...
22:01:01 <oerjan> yep, definitely exists
22:01:02 <CakeProphet> AnMaster: but I guess thought experiments are pretty bad jokes. BAM.
22:01:13 <Phantom_Hoover> Oh, so biject(1, 1) and biject(1.1, 1.1) would give completely different results?
22:01:17 <AnMaster> CakeProphet, well, I thought it was a bad joke
22:01:40 <AnMaster> CakeProphet, since I couldn't make sense of it as anything esle
22:01:54 <coppro> oerjan: yes, it exists, and it hurts my brain
22:02:05 <oerjan> http://people.math.gatech.edu/~berglund/Rubik/index.html
22:03:51 <CakeProphet> actually 6D is easier for me to imagine than 4D... because I can just picture two 3D spaces.
22:04:04 <AnMaster> CakeProphet, I can't imagine that either
22:04:14 <oerjan> Phantom_Hoover: the bijection cannot be a continuous function, this is part of what is known as Brouwer's theorem of Invariance of domain http://en.wikipedia.org/wiki/Invariance_of_domain
22:05:15 <Phantom_Hoover> In any case, this implies that 4-space can be embedded in 3-space.
22:06:31 <cpressey> It's the smallest possible space for us to have the perception of it that we do.
22:06:41 <Phantom_Hoover> CakeProphet, Stephen Hawking said something about that.
22:06:54 <AnMaster> cpressey, the same could be said for 4D if we lived in 4D
22:07:10 <AnMaster> Phantom_Hoover, so string theory is false?
22:07:14 <cpressey> AnMaster: For all we know, we do live in 4D. We just can't see it.
22:07:39 <cpressey> String theory is not just false, it's a load of hooey.
22:07:42 <CakeProphet> the fourth dimension is an awesome-to-fail gradient.
22:07:48 <AnMaster> Phantom_Hoover, which isn't an accurate desc
22:08:01 <oerjan> Phantom_Hoover: bijecting R^2 with R is easy, if you ignore the 0.99999... = 1.0000... issue (and that's easy to fix up afterwards). just interleave the bits of the two numbers. oh and sign, but that's also easy to fix.
22:08:18 <Phantom_Hoover> If there are 4 conventional space dimensions, it becomes an inverse cube law, and orbital deviations are magnified hugely.
22:08:30 <CakeProphet> cpressey: I think string theory is just overzealous in its categorization of physical phenomenon as dimensions
22:08:50 <CakeProphet> unfortunately, to get anywhere in physics these days you have to pretty much submit to string theory or you will not be taken seriously.
22:08:50 <oerjan> Phantom_Hoover: more _generally_, it's a theorem of set theory that the axiom of choice is equivalent to every infinite set M having a bijection with M^2
22:11:16 <CakeProphet> Is there a name for this magical (^) :: Set -> Integer -> Space operation?
22:11:42 <cpressey> Oh so THAT is what the Axiom of Choice is.
22:12:20 <CakeProphet> ah so... Set -> Integer -> DependentTypeHereLolHaskell
22:13:37 <oerjan> CakeProphet: if you identify an integer with a set having that many elements (as e.g. von Neumann numerals do), then A^B is just the set of functions from B to A
22:14:46 <CakeProphet> I think I understand, but as a consequence my brain kind of malfunctioned.
22:15:43 <oerjan> CakeProphet: this is also compatible with the cardinalities of the sets. e.g. the set of functions from B to A has |A|^|B| elements, where |A| and |B| are the cardinalities of A and B
22:16:02 <CakeProphet> from a set with two elements to R I guess.
22:16:02 <oerjan> so it's an intuitive notation
22:16:22 <Phantom_Hoover> Wait, doesn't 2 have to be a set of 2 elements each of which are in R?
22:17:02 <oerjan> Phantom_Hoover: no, it doesn't matter which set you use to represent 2 for this, as long as it has 2 elements
22:17:45 <oerjan> von Neumann numerals use {0,1}, defined recursively
22:18:09 <oerjan> Phantom_Hoover: it maps 0 to the first tuple element, and 1 to the second.
22:18:25 <oerjan> some might like to use {1,2} instead for this :D
22:18:46 <oerjan> more generally, an n-tuple in math is just a function from {1,...,n} to some set
22:20:04 <CakeProphet> hey, need to provide validity to your theory? Define it in terms of other theories!
22:20:23 <cpressey> I firmly believe AC to be independent of AC.
22:20:56 <AnMaster> oerjan, so that means there are just two types of infinite?
22:21:04 <AnMaster> oerjan, countable and uncountable?
22:21:31 <AnMaster> if that is true countable == uncountable which is not true afaik?
22:22:08 <oerjan> AnMaster: um no, there is an unending (in fact itself uncountable) collection. 2^M has larger size than M, always (Cantor diagonalization)
22:22:26 <oerjan> however M^2 has the same size as M if M is infinite
22:23:17 <oerjan> interestingly, there are infinitely many cardinalities a such that a^(aleph_0) = a and infinitely many such that this is false
22:23:42 * Phantom_Hoover wonders what a 3-cube bijected into 3-space would look like.
22:23:49 <oerjan> (any cardinality of the form 2^a for infinite a satisfies the equation, but limits need not
22:24:25 <oerjan> s/infinitely/uncountable/, really
22:25:38 <cpressey> Sounds strangely like relativizing proofs on oracle machines.
22:26:00 <oerjan> cpressey: well it's all diagonalization
22:26:12 <cpressey> "infinitely many [oracles] such that [P=NP] and infinitely many such that this is false"
22:27:08 <cpressey> Well, I don't *know* that there are inf. many -- the original result was "there exists at least one" -- but it seems quite reasonable
22:27:36 * Sgeo is in a dark cave with dripping water and weird violet crystals everywhere
22:27:44 <oerjan> well you can probably construct infinitely many oracles with the same power as a given one, at least
22:29:12 -!- calamari has joined.
22:29:34 <AnMaster> oerjan, "<oerjan> Phantom_Hoover: more _generally_, it's a theorem of set theory that the axiom of choice is equivalent to every infinite set M having a bijection with M^2" still confuse me then
22:29:37 <oerjan> Phantom_Hoover: 3-cube into 3-space? well if it's a cube without the borders, just use a suitable scaled tan function on each coordinate
22:29:42 <AnMaster> oerjan, every doesn't mean every?
22:29:48 <calamari> according to wikipedia: PostScript is a Turing-complete programming language, belonging to the concatenative group. PostScript is an interpreted, stack-based language similar to Forth but with strong dynamic typing, data structures inspired by those found in Lisp, scoped memory and, since language level 2, garbage collection
22:30:05 <AnMaster> oerjan, or that you have bijections between different cardinalities?
22:30:08 <calamari> this is just asking for a brainfuck port..
22:30:16 <oerjan> AnMaster: it has to be the same M in M and M^2
22:30:19 <AnMaster> oerjan, I thought bijection defined cardinality?
22:31:28 <cpressey> calamari: You know, I'm not sure anymore... I'm ok I guess. :) How're things with you?
22:31:36 <oerjan> calamari: now make it output a graphic representation of the bf run :D
22:32:34 * oerjan did a mandelbrot in postscript once. only black and white though, so not so impressive.
22:33:02 <oerjan> i have a vague idea that it _might_ have taken a while to get through the printer :D
22:36:07 <calamari> cpressey, I'm doing alright.. did something happen?
22:38:41 <cpressey> calamari: It's a long story. I've been bouncing around the USA the past two years, for bad reasons. Ended up near Chicago... will hopefully stay here for a while. But we'll see.
22:40:08 <cpressey> Yep, just a US permanent resident, because I married an American :)
22:40:49 <oerjan> CakeProphet: i've been assuming it's pronounced like the last part of "quintuple".
22:40:50 <AnMaster> hm, I wonder what an image that has a small depth of focus but has more than one such would look like
22:41:26 <CakeProphet> I suppose it's because the words that end in -tuple pronounce it both ways. right?
22:41:39 -!- Phantom_Hoover has quit (Remote host closed the connection).
22:41:41 <calamari> AnMaster, like a pinhole camera?
22:41:55 <calamari> or are you saying the opposite of that?
22:42:10 <AnMaster> calamari, like focused at one meter, then unfocused a bit, then focused again at about 10 meters
22:42:16 <calamari> well with a small enough pinhole, everything is in focus
22:42:25 <AnMaster> calamari, so completely different
22:44:14 <AnMaster> calamari, so 1 = completely in focus -1 = very out of focus
22:44:25 <AnMaster> with a period of half a meter between crossing the zero
22:44:28 <calamari> you might be able to (ab)use a raytracer into doing it
22:45:22 -!- ehirdiphone has joined.
22:45:31 -!- ehirdiphone has quit (Client Quit).
22:45:44 <oerjan> so close and yet so far
22:45:53 -!- ehirdiphone has joined.
22:47:36 <calamari> Chris: but yeah anyhow.. I am a divorced dad of 2.. hopefully you didn't end up in such a situation
22:47:56 <cpressey> calamari: I'm sorry to hear that...
22:48:00 <Sgeo> It feels so good to write out what the code should do before writing the code
22:48:59 <CakeProphet> hmmm... I bet you could extend Parsec to match/fail on any sort of pattern regardless of type.
22:49:14 <CakeProphet> and also fuzzily. To implement pattern matching.
22:49:32 <Sgeo> ehirdiphone, "How are you?"
22:49:37 <calamari> ehirdiphone, hehe.. mostly me not minding my own business :)
22:52:28 <AnMaster> calamari, wrt raytracer I mean
22:52:56 <AnMaster> calamari, still I wouldn't know how to make such an optics model
22:53:24 <cpressey> ehirdiphone: Me volunteering the fact that I've been galavanting around the US because I married an American.
22:54:32 <cpressey> Well, when enough crap starts happening to you, you start learning to count your blessings I guess :)
22:54:55 <calamari> AnMaster, I guess I'm thinking along the lines of modifying the source to mess with the smooth focus falloff and reset it at some point.. a dirty hack
22:55:28 <calamari> so when distance gets to 10 meters.. treat it like distance x-10 again
22:55:51 <AnMaster> calamari, well I guess one would have to find it in the source first
22:56:31 <ehirdiphone> I have a good shot at first or second shitlife ranking here methinks.
22:56:40 <calamari> yeah fmod.. wow, never heard of fmod until now.. cool stuff :)
22:56:51 <ehirdiphone> Does it get much worse than being wrongly institutionalised?
22:57:12 <cpressey> ehirdiphone: Yes. Yes, I for one would not deny you that trophy.
22:58:29 <calamari> for me, I think it was the sudden drop in standard of living that was jarring, but that's been a long time now so it's just life again hehe
22:58:31 <ehirdiphone> Wow, calamari invented modular arithmetic. Astonishing :P
22:59:23 <ehirdiphone> fmod is just the remainder operation generalised to reals
22:59:32 <calamari> maybe discovered :) definitely not invented
23:00:28 <calamari> I love being an argumentative ass.. just ask Gregor
23:00:43 <ehirdiphone> Actually fmod doesn't make a whole lot of sense really since modulo is remainder of *integer* division
23:01:16 <ehirdiphone> But it's a useful implementation over the reals of a common use-case of modular arithmetic.
23:02:42 <ehirdiphone> cpressey: So if we get gay married I could escape? Hey, it's a viable plan!
23:02:58 <ehirdiphone> ATTN anyone in America: I am free for marriage
23:03:18 <cpressey> ehirdiphone: Er, no can do -- there are laws against bigamy here.
23:03:19 <coppro> ehirdiphone: have fun consummating
23:04:26 <CakeProphet> I guess in C++ you coud define tuple generics.
23:04:32 <coppro> ehirdiphone: someone's challenging those laws' constitutionality I think
23:04:37 <coppro> CakeProphet: C++0x is getting them
23:04:43 <ehirdiphone> CakeProphet: c is a glorified pdp11 assembler
23:04:52 <Sgeo> ATTN any non-married person, then?
23:05:26 <ehirdiphone> CakeProphet: c++ is a functional language compiling down to a glorified object-oriented pdp11 assembler
23:06:06 <coppro> ehirdiphone: bigamy. Gay marriage has been legal here for several years
23:06:11 <calamari> ehirdiphone, but god hates fags </troll>
23:06:38 <ehirdiphone> coppro: the authors of the Canadian constitution were gay polyamorists
23:07:04 <calamari> Sgeo, sorry, arizona hates gays too
23:07:18 * calamari currently lives in the most hated state, lol
23:07:32 <coppro> there was never time for an actual SCC ruling on the legality of law banning gay marriage, IIRC; the government made it legal before that could happen. The closest it came when was Alberta (surprise, surprise) tried to legislate it and it was found to be outside provincial jurisdiction.
23:07:39 * CakeProphet is in Georgia. Doesn't get much more bigoted.
23:08:02 <calamari> CakeProphet, but nobody talks about that.. Arizona is in the spotlight
23:08:18 <coppro> I don't even get why everyone cares about Arizona
23:08:30 <ehirdiphone> CakeProphet: Georgia birthed Neutral Milk Hotel which makes up for it sucking.
23:08:31 <coppro> all they did is basically tell police officers to enforce the law
23:08:44 <CakeProphet> I don't get why everyone cares about the US.
23:08:56 <calamari> well from a racism perspective, I guess I don't really care.. however requiring me to carry my papers.. I care about that
23:09:06 <CakeProphet> ehirdiphone: ha. And a lot of other good musicians. I didn't know NMH was from Georgia though.
23:09:06 <ehirdiphone> CakeProphet: I have a friend in Georgia; say hi.
23:09:22 -!- MigoMipo has quit (Read error: Connection reset by peer).
23:10:01 -!- tombom has quit (Quit: Leaving).
23:10:10 <CakeProphet> hmmm, is singular-for-band-names a US thing?
23:10:57 <coppro> it's not significantly different from referring to a corporation by a singular name
23:11:08 <CakeProphet> so do you say "the class are dismissed." ???
23:11:09 <ehirdiphone> Pink Floyd were a band. Microsoft were ruled against in an antitrust case.
23:11:23 <oerjan> coppro: so basically it's the reason why US corporations are persons?
23:12:16 <ehirdiphone> When we hear "the class" we hear every member of that class.
23:12:18 <oerjan> ehirdiphone: isn't it a little early for christmas
23:12:30 <ehirdiphone> But it sounds awkward for a teacher to say that.
23:13:03 <CakeProphet> well, I presume because there is only one class.
23:13:27 <calamari> the class is dismissed ... imo :)
23:13:31 <CakeProphet> the classes are dismissed. The class is dismissed. That's how it works here at least.
23:13:43 <oerjan> CakeProphet: but does it own the means of production?
23:14:43 <ehirdiphone> Corporations as their members is a very socialist quirk :P
23:14:47 <CakeProphet> ehirdiphone: I hope class just so happens to be an implementation of a hash table...
23:14:54 <CakeProphet> because then we can implement a subset of lua.
23:15:23 <ehirdiphone> Well we need the Class class because every class must be an instance of it.
23:15:47 <coppro> I HEREBY FLAW YOUR LOGIC
23:16:39 <CakeProphet> Why would you even go through the trouble of defining a better C++
23:17:12 <cpressey> CakeProphet: Yes. Class must be a subclass of HashTable.
23:17:39 <CakeProphet> because if there are tuples... then they should immediately change every stdlib function that takes pointer arguments to set multiple return values.
23:18:08 <coppro> C library compatibility says no
23:19:49 <olsner> I will take some mana for this, you know... massive amounts of mana
23:20:05 <CakeProphet> also... why on earth would you call it C++0X
23:20:06 * coppro steals the mana and casts Gleemax
23:20:19 <pikhq> coppro: People are upset about Arizona because the law in question makes the police call out people who "look illegal" and demand proof that they are legal. Said proof being a birth certificate or naturalisation certificate.
23:20:24 <coppro> CakeProphet: It's a working name evolved from the fact that it was expected in the year '0X, where X is unknown
23:20:36 <CakeProphet> dude check out my language it's called $##E2
23:20:39 <cpressey> CakeProphet: Obviously, it is an esolang in disguise.
23:21:00 <pikhq> And these certificates are neither commonly carried around nor something that police should (IMO) have the authority to demand.
23:21:04 <ehirdiphone> CakeProphet: It's official C++ not some other project.
23:21:13 <olsner> c++1x, in which C++ standardises brainfuck as an alternate macro language
23:21:26 <coppro> *It's required that foreigners carry ID already. IIRC, it adds no requirement for citizens
23:22:19 <pikhq> coppro: Yes, foreigners are required to carry a passport or a green card. However, the law makes demands of *proof of citizenship*, not just proof of legal non-citizenship.
23:22:35 <coppro> oh, I didn't know about that
23:22:48 <coppro> I don't recall seeing that when I scanned the bill
23:22:55 <pikhq> And it also demands that they do this check on anyone who *looks illegal*.
23:23:15 <olsner> anyway, about defining a better C++ the thing I think I've realized that you pretty much can't make a better C++ but you can probably replace it with something else
23:23:18 <CakeProphet> isn't that what cops do anyways? not that it's justified...
23:23:19 <pikhq> Uh, yeah... Most people here are either immigrants or decendents of immigrants. Each and every person could be illegal.
23:23:19 <coppro> I believe the term is "reasonable suspicion:
23:23:30 <coppro> ehirdiphone: I looked this one up specifically
23:23:53 <ehirdiphone> coppro is defending the crazy right wing burkina
23:23:54 <pikhq> Also: this is a state with fucking Sherrif Arpaio in power.
23:24:06 <cpressey> Proof of citizenship? So permanent residents don't count? Jeez/
23:24:27 <pikhq> cpressey: Citizenship or legal residence.
23:24:33 <oerjan> ehirdiphone: it's just going through a faso
23:24:43 <calamari> pikhq, so glad I don't with in maricopa county... inmnocent into proven guilty throw right out the window with those prison camps
23:25:16 <calamari> wow, tons of typos in that.. just parse it :)
23:25:20 <pikhq> But, yeah. Arpaio would inevitably enforce that shit as "if e's not white or black, pull im the hell over."
23:25:33 <pikhq> (black so that he can claim not to be racist)
23:25:41 <pikhq> oerjan: No, but innocent is.
23:25:48 <ehirdiphone> pikhq: Yeah... Arpaio would totally use Spivak pronouns.
23:25:58 <pikhq> ehirdiphone: Totally. :P
23:26:00 <calamari> oerjan, that's what happens when I am reading a webpage and typing at the same time lol
23:26:14 <coppro> pikhq: also, you know the other part, right? That the police officer already has to be talking to the person for some violation?
23:27:08 <olsner> where C++ wins today there isn't really that much to compete with it I think... C is often too stupid/simplistic and everything else popular is too dynamic or slow or generally far-from-the-metal (like ... java and everything) or scary (like haskell)... Ada isn't bad at all really but everyone mistrusts it and it's a tad verbose and bondage&discipline
23:27:23 <pikhq> Though this in practice means a more asshole-type police officer will make up a reason to pull the person over.
23:27:27 <calamari> coppro, and you know that if a police officer follows you long enough in a car you WILL violate some traffic law sooner or later.
23:27:29 <ehirdiphone> Just going to toss in a crazy radical idea to turn this into a flamewar: Completely free immigration to and from any country.
23:27:43 <cpressey> Sadly, I need to be off, before I crack up. Bye all, have a pleasant evening.
23:27:52 <pikhq> ehirdiphone: My only question is whether or not it would be sane to implement this immediately.
23:27:52 -!- cpressey has quit (Quit: Arpaio would totally use Spivak pronouns.).
23:27:56 <coppro> calamari: sure. But the new law doesn't change that aspect of things significantly
23:28:14 <pikhq> Perhaps have it completely free immigration to and from the developed world and slowly add developing nations.
23:28:44 <ehirdiphone> pikhq: Yeah, it's not practical right know.
23:28:48 <pikhq> (so as to avoid a massive flood into developed nations)
23:28:52 <coppro> if the police officer thinks someone is an illegal immigrant, they can still tail you all day, get you for a violation, and ask you for documentation that you're allowed in the country
23:29:10 <ehirdiphone> But it'd make countries *compete*, thus yielding better countries,
23:29:22 <ehirdiphone> thus yielding better spread of population.
23:29:24 <pikhq> But, other than doubts about implementation particulars, I am entirely in favor of such a thing.
23:30:34 <pikhq> There is absolutely no sense in there not being free immigration between, say, Europe, USA, Canada, and Japan. None at all.
23:31:06 <pikhq> I can't even think of a *reason* for it beyond "DEM FOREIGNERS ARE EVIL TAKING AR JORBS".
23:31:11 <ehirdiphone> pikhq: BUT DOOD PPL ONLY WANT TO GO TO USA AND UK LOL
23:31:41 <oerjan> ehirdiphone: you're not supposed to disappear until after you've _finished_ writing Candlejack, you stu
23:32:08 <pikhq> Oh, also. "TERRORISTS".
23:32:14 <ehirdiphone> oerjan: olsner stopped me finishing Candlejack's name, you fo
23:32:24 <calamari> omg they took my job I can't work at mcdonalds anymore :(
23:32:46 <pikhq> Because, of course, Canada is a magical land made entirely of people who want to commit terrorism against the US.
23:33:04 <pikhq> But can't be assed to walk across the mostly unpatrolled border to do so.
23:33:39 <ehirdiphone> *recites the rest of the end of the world*
23:33:43 <pikhq> Why we have the border mostly unpatrolled *but* massive security checks on legal crossings is beyond me.
23:35:05 <ehirdiphone> Canada is awesome because it has European laws plus English (ok so I am biased) plus an awesome range of climates.
23:35:58 <pikhq> And its culture is like America, without all of the stupid. :P
23:36:18 <pikhq> Hrm. Actually, that makes it radically different; stupid is one of the defining attributes of American culture.
23:36:42 <ehirdiphone> pikhq: wrt terrorism: with the English skills of the average terrorist fearmonger, we will soon all be incredibly worried about planes being hijacked by museli
23:37:00 <ehirdiphone> OH GOD WITH THE ALMONDS AND THE OATS IN THE COCKPIT
23:37:01 -!- micahjohnston has joined.
23:37:29 <CakeProphet> would be nifty if Haskell had recursive types
23:37:34 <coppro> Nous avons aussi Français.
23:37:56 -!- ehirdiphone has quit (Quit: Get Colloquy for iPhone! http://mobile.colloquy.info).
23:38:17 -!- ehirdiphone has joined.
23:38:21 <CakeProphet> type Computation a b = (a ->(a, Computation))
23:38:42 <ehirdiphone> It's like Canada plus LAME FRENCHNESS and EXCESS COLD.
23:38:53 <oerjan> CakeProphet: the argument i hear for not allowing them is that it would make a _lot_ of errors accidentally well-typed
23:39:11 <ehirdiphone> CakeProphet: You missed some type arguments...
23:39:39 <CakeProphet> type Computation a b = (a ->(b, Computation))
23:40:04 <oerjan> ehirdiphone: well not memory safety
23:40:50 <oerjan> ehirdiphone: it cannot cause it to call a function with something it cannot handle, only make an unexpected infinite loop
23:41:02 <coppro> C'est à Ottawa et un peu de Nouveau-Brunswick.
23:41:47 <ehirdiphone> Ayeight mayb I starten tak'n i'scrut'bl 'nvented Enggish dylect.
23:43:01 <pikhq> ehirdiphone: Ai shinkku Ai amu goingu tsu taruku in an oddo daiarekuto maiserufu.
23:43:32 <oerjan> ehirdiphone: there's nothing you can do with recursive types that you cannot already simulate with newtypes, afaik
23:44:01 <coppro> pikhq: nice bastardization
23:44:09 <coppro> ehirdiphone's is just bad
23:44:09 <oerjan> (and they compile to identical machine code, presumably)
23:44:12 <pikhq> micahjohnston: Korya "I think I am going to talk in an odd dialect myself.", nihongoppoi hatsuon de.
23:44:50 <pikhq> (Watashi ha ima, IME ga nai no de, roomaji wo tsukaeru. Gomen ne.)
23:44:56 <micahjohnston> pikhq: wakarimashita demo nihongo sukoshi wakarimasu
23:45:08 <micahjohnston> I don't even know if what I just said is right at all
23:45:27 <pikhq> micahjohnston: Quite correct, though oddly formal.
23:46:09 <ehirdiphone> Desu Tokyo Nihongo desu Nihongo dedu desu.
23:46:22 <oerjan> pikhq: e syns japansk e å gå før langt
23:46:38 <AnMaster> oerjan, aha, så det är japanska?
23:46:59 <pikhq> ehirdiphone: This is about on par with "Is London English is English id is." :P
23:47:30 <oerjan> AnMaster: well some of it at least
23:47:35 <pikhq> "Is is; Edinburgh is English."
23:48:03 <olsner> pikhq: you know japanese? ehird's line read like "I understood but I understand a little japanese" to me... which seems to be missing some kind of contrast to motivate the 'but'
23:48:05 <pikhq> micahjohnston: He's citing a bizarre meme, not actual Japanese.
23:48:17 <pikhq> olsner: I do, in fact, know Japanese.
23:48:28 <ehirdiphone> Wo desu Nihongo gaijin desu wo wo desUsaka
23:48:32 <pikhq> ehirdiphone: Yeah, but it's *very* uncommon.
23:48:44 <coppro> ehirdiphone, on the other hand, does not speak a lick of Japanese
23:49:07 <AnMaster> olsner, så länge de pratar japanska föreslår jag att vi pratar "skandinaviska"
23:49:25 <olsner> japansk e då ikkje skickelig vansklig, bare *litt* vansklig
23:49:29 <pikhq> micahjohnston: "Watashi ha wakuru, kedo chotto dake nihongo wo wakaru."
23:49:50 <olsner> AnMaster: nån sorts försök till det
23:50:12 <pikhq> ehirdiphone: "The is English foreign is the the is?eh."
23:50:21 * coppro is getting annoyed that Google translation can't do Romanji (this is not an invitation to move to Kanji, thanks)
23:50:21 <olsner> CakeProphet: gobbledygook time! join in!
23:50:29 <CakeProphet> isn't this is an only-speak-American IRC server?
23:50:38 <pikhq> coppro: I don't have an IME, so I am not using my uber kanji-knowledge.
23:50:57 <oerjan> AnMaster: norsk e jo så lætt, e har snakkæ dæ si e va lit'n glønjtonge
23:50:58 <pikhq> (I can write all the jouyou kanji. I know all the readings for about half of them.)
23:50:59 <AnMaster> CakeProphet, not american. UK English
23:51:17 <AnMaster> oerjan, vad är "snakkæ" och "glønjtonge"?
23:51:37 <oerjan> AnMaster: "spoken" and "boy child"
23:51:43 <pikhq> coppro: You should fix that.
23:51:45 <ehirdiphone> Fvvjcogctpdtpsyuvug h hpctptxtc bgpdyxtu uuuuuuäfghtįåæ.
23:51:52 <pikhq> If nothing else, it lets you write kanji in English. :P
23:51:55 <AnMaster> oerjan, huh, doesn't make more sense even with that
23:52:03 <coppro> pikhq: Possibly. I want to brush up on French first before I attack another language
23:52:14 <oerjan> AnMaster: it's dialect of course :D
23:52:21 <pikhq> (my notes are now incredibly hard to read unless you're bilingual. And can read an invented script of mine.)
23:52:45 <olsner> snakkae == snacka, såklart
23:52:48 <AnMaster> oerjan, the Swedish dialects I can do passably aren't even half as evil
23:53:00 <coppro> pikhq: Though I do have some questions on pronunciation of Japanese, if we can move to /msg
23:53:13 <ehirdiphone> I am not about to start speaking in Geordie.
23:53:22 <AnMaster> oerjan, now I wish I knew dalmål
23:54:16 <pikhq> elliottable: o kamitachi
23:54:17 <AnMaster> Gregor, anything screenshot-worthy yet?
23:54:35 <olsner> I know some greek! mia nihta mono den ftani
23:54:50 <AnMaster> olsner, I read that last word as "fthang" first XD
23:55:27 <olsner> yeah, I thought it was cthulhu-speak for a long while, until I managed to find a forum post identifying it as greek
23:56:17 <olsner> (the name of a Shulman song, btw, http://open.spotify.com/track/77srFu8ZmyivyhM9YZkNlk)
23:56:33 <oerjan> cthulhu-speak is easy once you get the fthang of it
23:57:13 <olsner> AnMaster: sucks to be you then
23:57:26 <AnMaster> olsner, as far as I knew there was no linux client for non-premium?
23:57:37 <olsner> well, it's bound to be on your favourite p2p network anyway
23:57:58 <Sgeo> WINE isn't perfect
23:58:04 <AnMaster> olsner, I don't deal in that. I tend to order my music from naxosdirect after hearing it on P2 :P
23:58:19 <olsner> spotify works fine in Wine anyway, except for playback of local files
23:58:19 <AnMaster> Sgeo, iirc the spotify client doesn't work well at all with wine
23:59:12 <ehirdiphone> LET'S DO SOME CANNABIS DRUGS THEN DEAL IN COPYRIGHT-INFRINGING MATERIALS
23:59:15 <olsner> I've had no problems... but maybe you just need to uninstall pulseaudio to make it work
23:59:26 * Sgeo would love to use ReactOS as his primary OS
23:59:29 <AnMaster> olsner, I don't use pulseaudio on that system
23:59:57 <olsner> bah! why do they even bother inventing this crap? to make sound more difficult?