00:14:24 -!- copumpkin has joined.
00:16:26 -!- pumpkin has quit (Ping timeout: 250 seconds).
00:18:53 -!- copumpkin has quit (Ping timeout: 246 seconds).
00:37:19 <Gregor> Damn, why can't gelfload load normal binaries :(
00:40:24 <Gregor> I should make gelfload a library so you could make platform binaries with platform-independent plugins :)
00:42:00 <pikhq_> Why the heck are you working on gelfload again? :P
00:42:08 <Gregor> Because I have libdl.so X-P
00:42:25 <Gregor> gelfload as a library is like portlibdl :)
00:42:44 <pikhq_> Except you have to deal with nasty things like "calling conventions".
00:45:08 <Gregor> Those problems are MOSTLY blown out of proportion :P
00:47:14 <Gregor> Floats are particularly nasty :P
00:47:17 <pikhq_> Many architectures have a single calling convention or a single one that's actually common...
00:47:21 <pikhq_> And then we get x86 and x86_64.
00:47:34 -!- BeholdMyGlory has quit (Remote host closed the connection).
00:48:05 <pikhq_> x86 has, lessee. The set of calling conventions used by Windows is comprehensive, isn't it?
00:48:51 <Gregor> Which is quite retarded.
00:49:00 <Gregor> cdecl is MOSTLY the same as Unix though.
00:49:05 <Gregor> http://www.agner.org/optimize/calling_conventions.pdf <-- so awesome
00:49:08 <pikhq_> x86_64 has the Microsoft calling convention and the standard one.
00:50:00 -!- augur has quit (Remote host closed the connection).
00:50:04 <pikhq_> I should here note that I only really refer to C calling conventions.
00:50:14 <pikhq_> C++? Abandon all hope, ye who enter here.
00:50:53 <Gregor> Well yeah, obviously C++ is hell.
00:51:41 <Gregor> Anyway, I think libdl.so would be a pretty awesome homepage for gelfload and connected projects :P
00:51:42 <pikhq_> Perhaps the only saving grace there is GCC has a single C++ calling convetion.
00:52:28 <Gregor> Anyway, I was thinking about hacking up gelfload to do "approximation" so you could load old binaries and such if things that changed names just so happened to remain mostly compatible (e.g. libc version hell)
00:52:45 <Gregor> But for some reason I can't load native binaries even when I load all deps with dlopen ...
00:53:13 <pikhq_> Strange; I seem to recall using gelfload on arbitrary binaries just fine.
00:53:32 <Gregor> Arbitrary ... statically compiled binaries? :P
00:54:42 * pikhq_ should actually build this GCC 4.6.0 RC and binutils 2.21.51...
00:54:55 <pikhq_> And let the LTO make static compilation suck less!
00:55:22 -!- cal153 has quit (Ping timeout: 255 seconds).
00:57:06 <Gregor> lol, this seems to have failed in exit() :P
00:57:10 <Gregor> WURVE when that happens :P
00:57:19 <Gregor> It's so easy to make that fail in a runtime ELF loader.
00:57:45 <pikhq_> (GCC 4.6.0 has LTO not suck, and binutils 2.21.51 has linker plugin support in a genuinely stable linker)
00:57:47 <Gregor> Oh, I'll bet it's because it double-exits.
00:59:34 <Gregor> WHAT IN GODS NAME IS GOING ON
01:01:07 <Gregor> http://sprunge.us/OPHg
01:03:52 <Gregor> So I/O seems a bit ... "odd"
01:03:59 <Gregor> But /bin/ls works modulo segfault at exit
01:04:16 <Gregor> And xterm works which is kinda awesome :)
01:05:11 <Gregor> lolwtf I can launch gdb :P
01:05:29 <Gregor> $ ./src/gelfload ./src/gelfload ./src/gelfload # INCEPTION
01:14:14 -!- cheater00 has joined.
01:17:25 -!- cheater99 has quit (Ping timeout: 255 seconds).
01:19:59 -!- zzo38 has joined.
01:24:21 -!- cal153 has joined.
01:36:01 -!- augur has joined.
01:36:03 -!- HackEgo has quit (Ping timeout: 276 seconds).
01:36:17 -!- EgoBot has quit (Ping timeout: 252 seconds).
01:38:47 -!- copumpkin has joined.
01:42:02 <Gregor> Found /lib/libc.so.6 for libc.so.5 :)
01:42:17 <zzo38> Are the files compatible?
01:43:08 <Gregor> Depends on what you use.
01:43:30 <Gregor> Anybody have an x86_64 BSD system floating about they'd like to throw me a binary from?
01:43:50 <Gregor> Or maybe a simpler /usr/bin/yes?
01:48:06 -!- hagb4rd has joined.
01:56:47 -!- wareya_ has joined.
01:58:08 <Gregor> Mmmmmmmmmmmmmmmm corned beef and cabbage
01:59:25 -!- wareya has quit (Read error: Operation timed out).
02:07:17 -!- sebbu2 has joined.
02:07:17 -!- sebbu2 has quit (Changing host).
02:07:17 -!- sebbu2 has joined.
02:07:42 <zzo38> I guess no BSDers.
02:08:58 <Gregor> Welp, time to make myself a FreeBSD install then!
02:10:29 -!- sebbu has quit (Ping timeout: 276 seconds).
02:28:23 <pikhq_> Mountain Dew Throwback is now a permanent member of the Pepsi line of products.
02:29:10 <pikhq_> Sucrose tastes so much better.
02:29:53 <pikhq_> It's really *quite* apparent in sodas.
02:30:18 <Gregor> You realize I make my own soda, right? :P
02:30:34 <pikhq_> You're also anosmic. :P
02:30:49 <Gregor> I ALSO have an ELF loader that can load itself :P
02:30:55 <pikhq_> Yes, and that is awesome.
02:32:02 <Gregor> Aww foo, can't run FreeBSD ls :(
02:32:36 <Gregor> Oh, interesting, it's in relocating, not running ...
02:33:06 * Sgeo installs Glary
02:33:15 <Sgeo> Ninite has it, CNEt's heard of it
02:34:49 <Gregor> Symbol undefined: 'atexit' wuh?
02:35:39 <Gregor> And I got Found /lib/libc.so.6 for libc.so.7
02:36:54 <Gregor> Oh, or is atexit maybe a macro in glibc?
02:36:58 <Sgeo> What's bad about UAC not dimming the desktop?
02:37:17 <Sgeo> It's not like this is Ubuntu, where a rogue application can just fake a dialog asking for the password...
02:37:57 <Gregor> Looking for atexit in libmetahost_libc.so.7: (nil) ... yuh
02:38:07 <Gregor> $ nm -D /lib/libc.so.6 | grep atexit
02:38:08 <Gregor> 0000000000036870 T __cxa_atexit
02:38:21 <Gregor> In PRINCIPLE I could run weirdo binaries, in PRACTICE I can't :P
02:38:39 <Gregor> Because the specs don't require certain things to be symbols and certain things to be macros.
02:39:30 <Sgeo> Microsoft apparently has heard of Glary utilities. Though maybe I should learn how signed binaries work on Windows
02:40:02 <Sgeo> Maybe they're not called signed binaries
02:40:15 <pikhq_> Gregor: Just make a "library" that consists of calls to each libc function.
02:40:45 <pikhq_> I of course mean "make a program that makes".
02:40:55 <Gregor> I'd have to make them ... from man pages or some such lunacy.
02:41:56 <Gregor> I think I'll make a generishims library.
02:43:29 <Sgeo> Ok, Glary is clearly demented
02:47:33 <Gregor> I'mma try with old Linux binaries instead.
02:56:57 <Gregor> ... in the eventuality that I download any :P
03:00:46 <pikhq_> Try getting ahold of some Loki games.
03:01:19 <pikhq_> (warning: they have more dependencies than just libc)
03:02:15 <pikhq_> Oh, SDL is still ABI compatible with what Loki wrote way back when.
03:05:04 <Gregor> Suggested old Lolki game?
03:05:17 <pikhq_> I dunno, Alpha Centauri?
03:06:34 * pikhq_ is actually surprised that Loki had any success porting games.
03:06:56 <pikhq_> They had 10 employees, and there was basically no infrastructure for doing games on Linux.
03:07:04 <pikhq_> SDL exists because they needed it.
03:07:38 <Gregor> Somehow this random guy got their Sim City 3000 port running on Fedora in 2010 ...
03:08:10 <pikhq_> It's actually quite commonly done.
03:08:22 <pikhq_> Linux is still system call compatible, you see.
03:08:46 <pikhq_> Just a matter of getting the entire set of libraries in place, and voila.
03:08:50 <Gregor> Oh, you think he bundled it with l---right
03:09:19 <pikhq_> IIRC, Gentoo has the appropriate libraries in emul-linux-loki-compat or some such.
03:11:40 <pikhq_> And, heck, even if it weren't for such convenient things, you could always just install a chroot of old Debian.
03:12:55 <pikhq_> Windows is still ABI compatible with Windows 1.0.
03:13:02 <pikhq_> Linux can't even keep ABI for 10 years.
03:13:54 <Gregor> Mac OS X isn't ABI-compatible with earlier versions of Mac OS X.
03:14:04 <Gregor> Also, 64-bit versions of Windows can't run Windows 1.0 binaries :P
03:14:04 <quintopia> compatibility was windows's number one goal for quite a while
03:14:20 <pikhq_> Ah, true, Macs are even worse off than Linux.
03:14:32 <pikhq_> It's still *possible* to run ancient binaries.
03:14:42 <Gregor> Fekk, potato binaries still work X-D
03:14:44 <pikhq_> Mac's had 2 ABI breaks so far.
03:15:12 <quintopia> apple is all about the "if it's broke, throw it out and replace it"
03:15:15 <pikhq_> Debian potato was on Linux libc, wasn't it?
03:15:26 <pikhq_> No, no, hamm was that transition.
03:15:48 <pikhq_> Yup, potato *non-C++* binaries will still work.
03:16:53 <pikhq_> (C++ ABI has had a breakage since then)
03:27:44 <Gregor> Call to undefined symbol __setfpucw
03:27:44 <Gregor> Call to undefined symbol __libc_init
03:27:44 <Gregor> Call to undefined symbol atexit
03:37:12 <Gregor> Program terminated with signal 11, Segmentation fault.
03:37:12 <Gregor> #0 0x0804c306 in ?? ()
03:43:32 <Gregor> 804c306: f6 44 43 01 40 testb $0x40,0x1(%ebx,%eax,2)
03:44:00 <quintopia> i'll see your "hmmmmmmm" and raise you a "hmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm"
03:44:25 <Gregor> Oh that's interesting, it's failing somewhere in the result of getopt_long_only ...
03:46:09 <Gregor> ... why is it dereferencing the result of getopt_long_only ...
03:46:39 <Gregor> Did getopt_long_only used to have a different API?
03:48:13 <Gregor> Oh, never mind, I misread.
03:49:05 <Gregor> That's odd, it's dragging something out of BSS ... that's zeroed still ...
03:53:15 -!- augur has quit (Remote host closed the connection).
03:54:32 <zzo38> In this game http://www.chessvariants.org/index/msdisplay.php?itemid=MSxiangqivsortho the inventor claims that there is no checkmate possible on first move. However, me and two other people say the Chinese side can checkmate immediately if they play first. Look at the picture (under "Setup"), it should be clear. What do *you* think?
03:55:45 <zzo38> Actually, *three* other people.
03:56:48 <Gregor> Seems like FreeBSD and GNU might have incompatible internal __mbrtowc functions :P
03:57:04 <zzo38> Gregor: What are __mbrtowc functions?
03:57:47 <Gregor> mbrtowc - convert a multibyte sequence to a wide character // essentially a UTF-8 to UCS-16 converter (in the common Unicode case)
03:58:00 <Gregor> But it's implemented as a macro on both FreeBSD and GNU.
03:58:15 <Gregor> To different versions of an internal __mbrtowc >_>
03:58:38 <Gregor> OHHEY! I just ran BSD pwd!
03:59:14 <zzo38> OK, now I know. Could you possibly make a kind of program patching for this?
03:59:57 <Gregor> zzo38: That's what I'm doing when there are functions that simply aren't supported on the other system, but it's harder when they're just incompatible >_>
04:04:08 <zzo38> How should I make the icon for the Courier piece in Courier Chess?
04:04:24 <Gregor> ... by finding someone who can draw? :P
04:04:24 <zzo38> (It is called "courier" and also "runner")
04:05:05 <zzo38> And my question is what kind of shape.
04:05:56 <Gregor> pikhq_: Stupidly, old (like, libc5 old) Linux binaries are being more difficult than FreeBSD binaries :P
04:06:30 <pikhq_> Gregor: BSD is better-written. ;)
04:07:35 <zzo38> Some of the pieces including FIDE chess, I just look at the SVG files in Wikipedia and typed in the same numbers. And for compound Archbishop/Marshal/Amazon, I take parts of it and make combined. Some pieces I look at other example and put my own numbers in. But some is difficult.
04:10:36 <zzo38> That is why I ask about Courier Chess.
04:39:03 -!- augur has joined.
04:47:13 <Sgeo> And, doing the recitation video's problem, I flat out forget that dot products are scalars
04:47:25 <Sgeo> I only remembered as the guy started speaking
04:48:28 <pikhq_> I am *very* disappointed at the limits on radiation.
04:49:13 <pikhq_> The legal limit for radiation level in irradiation is a bit too low to allow for *shelf-stable meat*.
04:52:09 <Sgeo> "For which angle θ is the component of A in the direction of B equal to 0."
04:52:26 <Sgeo> Is it bad if I don't do any math for that probem since the answer is so blatently obvious?
04:59:01 -!- augur has quit (Remote host closed the connection).
05:08:33 -!- sftp has quit (Remote host closed the connection).
05:08:47 <zzo38> I write some computer programs for Free Geek and make some changes to their other programs that I have been asked to do. I also help them make whatever documents and stuff they need in TeX, and make their logo in METAFONT.
05:11:18 -!- zzo38 has quit (Quit: This is a quit message!!! Now you have to play chess!!!).
05:12:59 <Sgeo> Solution sheet did not list theta=3pi/2 as an answer
05:13:03 <Sgeo> Who do I complain to?
05:13:47 <Sgeo> Then again, they also didn't list 2pi + pi/2 as an answer, and I wouldn't expect them to
05:20:28 -!- augur has joined.
05:20:46 -!- augur has quit (Remote host closed the connection).
05:29:59 * pikhq_ would like to beat everyone who thinks that English is and/or should be the sole language of the United States.
05:30:35 <pikhq_> Official languages of members of the United States include: English, Hawai'ian, Samoan, Chamorro, Carolinian, Spanish, French, and German.
05:31:28 <pikhq_> Though Finland could join the US, thereby solving that.
05:31:53 <pikhq_> Only really takes the consent of Finland & Congress.
05:32:16 <pikhq_> Or enough insanity to engage in a war of conquest.
05:33:28 <oklopol> would be pretty cool if we suddenly decided to invade usa, flew there and started beating ppl up
05:34:00 -!- augur has joined.
05:34:08 <pikhq_> One of the few quick fixes to our political system. :P
05:34:23 <oklopol> there's five million of us so if we went to new york, each of us would only have to beat up a few guys
05:34:49 <pikhq_> Man, just taking over New York would really fuck up the US.
05:35:06 <oklopol> i read somewhere really reliable that we have the second most weapons per guy in here. unfortunately you were the first.
05:35:20 <pikhq_> Not that it'd stop Congress from wanting to nuke New York after that.
05:35:34 <pikhq_> Good thing Congress doesn't have the power to nuke anything.
05:35:48 <pikhq_> *Unfortunately*, the President can nuke anything for any reason whatsoever at any time.
05:36:05 <oklopol> maybe that's the big change obama meant
05:36:07 -!- augur has quit (Remote host closed the connection).
05:36:21 <pikhq_> Not even joking about that, though.
05:36:34 <oklopol> well yeah that's what i've understood
05:36:41 <pikhq_> There is always one guy near the President with the equipment to signal a launch.
05:36:59 <oklopol> that is sorta hard to believe
05:38:02 <pikhq_> http://en.wikipedia.org/wiki/Nuclear_football Not fucking kidding.
05:57:09 -!- augur has joined.
05:57:18 -!- augur has quit (Remote host closed the connection).
06:03:56 -!- augur has joined.
06:19:11 -!- Slereah has quit (Ping timeout: 252 seconds).
06:20:28 -!- augur has quit (Remote host closed the connection).
06:22:16 <fizzie> Nuclear football, the sport of real men.
06:22:49 -!- wareya_ has changed nick to Alex_Megaroide.
06:25:24 -!- Slereah has joined.
06:35:59 -!- asiekierka has joined.
06:36:07 <quintopia> you mean european nuclear football, right?
06:36:29 <quintopia> you're referencing pikhq_'s comment
06:36:40 <quintopia> which is undoubtedly an article about the "briefcase"
06:55:15 <pikhq_> Which is called a football for stupid reasons.
07:00:07 <quintopia> presumably because it is meant to be protected by the person carrying it the way a runningback protects a football
07:02:54 <pikhq_> quintopia: No, because of media.
07:08:14 -!- variable has quit (Remote host closed the connection).
07:16:07 -!- asiekierka has quit (Ping timeout: 250 seconds).
07:46:03 * cheater00 is very happy because he's getting a free LYAH book
07:51:04 -!- Slereah has quit (Ping timeout: 240 seconds).
07:51:11 <Ilari> APNIC space fragmentation (total amount of non-reserved blocks of space smaller than): 1M: 1.786, 512k: 1.442, 256k: 0.895, 128k: 0.629, 64k: 0.329, 32k: 0.229, 16k: 0.154, 8k: 0.096, 4k: 0.062, 2k: 0.036, 1k: 0.024, 512: 0.017.
07:53:50 <Ilari> Yes, APNIC has over /14 (256k) worth of /24s. Allocating those would take fair amount of time, except that allocations start to seriously fragment when larger blocks are gone.
07:56:22 -!- augur has joined.
07:57:32 -!- Slereah has joined.
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:02:14 <Ilari> Just those small blocks will create something like 1500 fragments by the time APNIC depletes.
08:04:50 <Ilari> APNIC current maximum block size is 1M, so anything above that will be fragmented.
08:05:42 <Ilari> 2M would fragment into 2 blocks, 4M would fragment into 4, 8M would fragment into 11.
08:08:11 -!- Alex_Megaroide has changed nick to wareya.
08:18:45 -!- cheater99 has joined.
08:19:45 -!- cheater00 has quit (Ping timeout: 255 seconds).
08:26:49 <cheater99> Ilari: why would 8m fragment into 11?
08:28:24 <Ilari> There's only 5 1M blocks (and nothing larger non-reserved).
08:30:31 <Ilari> Large blocks are going to run out much faster than smaller ones, causing loads of fragmentation.
08:33:31 -!- Slereah has quit (Ping timeout: 246 seconds).
08:40:02 -!- Slereah has joined.
09:20:16 -!- augur has quit (Remote host closed the connection).
09:42:11 <cheater99> Ilari: can't they defragment them?
09:43:10 <cheater99> it would be perfectly imaginable people would be willing to swap blocks
09:43:45 <Ilari> That would mean renumbering, and renumbering is big amount of work, especially with IPv4.
09:44:05 -!- augur has joined.
09:45:23 <Ilari> With IPv6, renumbering is somewhat easier, and allocation strategies will also result much less renumbering as blocks grow.
09:47:41 <Ilari> Oh, and while APNIC will be source of something like only 1500 fragments, IP address transfers will result in loads more.
09:48:20 -!- cheater- has joined.
09:48:21 -!- cheater99 has quit (Ping timeout: 255 seconds).
10:44:02 <augur> http://satwcomic.com/how-to-keep-friends
11:23:08 -!- cheater99 has joined.
11:23:09 -!- cheater99 has left (?).
11:23:26 -!- cheater99 has joined.
11:24:36 -!- cheater- has quit (Ping timeout: 255 seconds).
11:42:46 <Ilari> According to latest stats file, APNIC has 35 202 304 IPv4 addresses available. That's barely over /7 worth of space (/6.93).
11:43:11 <cheater99> http://en.wikipedia.org/wiki/Dyson_number
11:49:22 -!- hagb4rd2 has joined.
11:52:29 -!- hagb4rd has quit (Ping timeout: 255 seconds).
11:53:30 -!- sebbu2 has changed nick to sebbu.
11:57:25 -!- copumpkin has quit (Ping timeout: 240 seconds).
11:57:54 -!- copumpkin has joined.
12:46:10 -!- FireFly has joined.
13:13:14 -!- Phantom_Hoover has joined.
13:20:26 -!- BeholdMyGlory has joined.
13:25:16 -!- cheater99 has quit (Read error: Connection reset by peer).
13:25:51 -!- hagb4rd2 has changed nick to hagb4rd.
13:26:31 -!- cheater99 has joined.
13:32:15 -!- cheater00 has joined.
13:34:08 -!- cheater99 has quit (Ping timeout: 255 seconds).
13:56:39 -!- BeholdMyGlory has quit (Remote host closed the connection).
13:57:35 -!- BeholdMyGlory has joined.
14:43:23 -!- Phantom_Hoover has quit (Quit: Leaving).
14:43:43 -!- Phantom_Hoover has joined.
14:44:54 -!- FireFly has quit (Quit: swatted to death).
14:51:34 <Gregor> $ ./src/gelfload /bin/bash
14:51:35 <Gregor> Symbol undefined: '__gmon_start__'
14:51:41 * Gregor was wondering why it wasn't working.
14:51:51 <Gregor> The reason: Oh, it was working perfectly, I'm just an idiot :P
14:52:08 <Gregor> (Turns out THAT'S WHAT BASH LOOKS LIKE DURP)
14:56:51 <Phantom_Hoover> Gregor, does this mean that Microcosm isn't actually utterly dead?
14:57:16 <Gregor> Phantom_Hoover: Microcosm is a project that I refuse to make entirely my project, so it is as alive as other people are willing to let it be :P
14:57:26 <Gregor> Phantom_Hoover: gelfload on the other hand is alive and well, but distinct from Microcosm.
14:57:45 <Phantom_Hoover> But it ground to a halt when you couldn't work out a VFS structure you liked, no?
14:58:08 -!- copumpkin has quit (Quit: Computer has gone to sleep.).
14:58:18 <Gregor> That's because everybody wanted ME to do the VFS, all I wanted was for there to BE a VFS.
14:58:43 <Gregor> I'd be happy for somebody else to design (and implement :P ) it, I just don't want to get stuck with direct FS, since that'll make Windows a lame duck.
14:59:51 <Gregor> And retarded limitations on file naming, not the best mapping of modes, etc.
15:04:02 -!- BeholdMyGlory has quit (Remote host closed the connection).
15:21:24 -!- asiekierka has joined.
15:21:54 <Phantom_Hoover> http://www.aolnews.com/2011/03/17/virginia-middle-school-students-suspended-for-oregano-possession/?test=latestnews
15:21:59 -!- asiekierka has quit (Remote host closed the connection).
15:22:05 -!- asiekierka has joined.
15:23:30 -!- MigoMipo has joined.
15:39:22 -!- cheater00 has quit (Read error: Connection reset by peer).
15:40:24 -!- cheater99 has joined.
15:41:16 <Gregor> Phantom_Hoover: However, Microcosm or otherwise, I do have this:
15:41:31 <Gregor> $ ./src/gelfload ./src/gelfload ./src/gelfload ./src/gelfload ./src/gelfload ./src/gelfload ./src/gelfload ./src/gelfload ./src/gelfload ./src/gelfload # INCEPTION
15:43:22 <Gregor> (Note that that only works by complete coincidence, btw :P
15:47:15 <Gregor> Well, gelfload only works because it's configured to be loaded by the host ELF (or whatever) loader into an area of memory that it's unlikely that the guest ELF will be loaded into.
15:47:37 <Gregor> It just so happens that when loading itself, the process of replacing that area of memory with ... well, itself is sufficiently atomic to not segfault.
15:48:09 <Phantom_Hoover> Gregor, so it's theoretically possible for gelfload to fail for no apparent reason due to unfortunate allocation?
15:48:31 <Gregor> Yes, but by that token it's theoretically possible for /lib/ld-linux.so to fail for the same reason.
15:48:57 <Gregor> Well, I could intentionally make a binary that would fail in that way :P
15:49:07 <Gregor> But binaries made for ELF/Linux know where ld-linux is.
15:49:19 <Gregor> They of course don't know where gelfload is.
15:51:27 -!- hagb4rd has changed nick to hagb4rd|afk.
15:54:44 <Gregor> http://codu.org/tmp/low.tar.bz2 <-- done
15:55:03 <Gregor> Apparently ld-linux is just barely smart enough to not load over itself, so it kills itself instead.
15:55:37 <Gregor> The result though is just "Killed" to stderr and $? == 137
15:57:26 <Ilari> Does it issue the signal to self or is that kernel killing process after exec goes sour too late to back it off?
15:57:40 <Gregor> The kernel has no idea.
15:57:45 <Gregor> It's totally userland.
15:58:05 <Gregor> The kernel doesn't even know how to load dynamic binaries.
15:58:06 <Ilari> strace should show it then?
15:58:37 <Gregor> strace can have ... unique behavior when things go wrong in the loader :P
15:59:01 -!- impomatic has joined.
15:59:32 <Ilari> APNIC down 0.26: 1k to Malaysia, 3x1M+512k+3x256k to China, 256 to India. 1.84 blocks remain.
16:00:15 <Ilari> Logaritmic size: /7.120
16:00:55 <Ilari> Relative allocation size: 12.4% (!)
16:01:54 <Ilari> Oh, and two of those 1M blocks were part of 2M block APNIC didn't have space to allocate in one block.
16:19:11 <Ilari> The existing January monthly record has already been slammed (and it has been only 18 days instead of full 31).
16:19:26 -!- Phantom_Hoover has quit (Quit: Leaving).
16:20:35 -!- Phantom_Hoover has joined.
16:22:12 <Vorpal> <Gregor> Apparently ld-linux is just barely smart enough to not load over itself, so it kills itself instead. <-- what are you doing for this to be of relevance?
16:22:26 <Ilari> I get 30-day figure of 2.34. Ouch.
16:24:15 <Vorpal> <Gregor> strace can have ... unique behavior when things go wrong in the loader :P <-- strange, after all it just ptraces system calls
16:24:16 <Ilari> At that rate, depletion in about 3.5 weeks.
16:24:17 <Sgeo> Phantom_Hoover, hmm?
16:24:54 -!- cheater00 has joined.
16:25:00 <Sgeo> Phantom_Hoover, I think Chrome has a thing where it submits inability to connect to some Google server, and if a lot of people can't connect...
16:25:10 <Vorpal> Ilari, when is depletion for RIPE?
16:25:34 <Phantom_Hoover> Sgeo, it couldn't connect to the page because *my WiFi was down*.
16:26:51 -!- cheater99 has quit (Ping timeout: 255 seconds).
16:28:40 <Phantom_Hoover> Gregor, stracing low shows it dying during the execve call.
16:29:01 <Ilari> Vorpal: Lagerholm (ipv4depletion.com) says 2012-10-17. Huston (potaroo.net) doesn't seem to give estimate until APNIC depletes (and RIPE becomes next). RIPE itself says "this year".
16:30:12 <Ilari> For some reason, Lagerholm gives really optimistic estimates for RIRs. Huston is much more pessimistic (might not be as pessimistic as reality). With IANA depletion, it was the other way around.
16:31:16 <Ilari> Also, ARIN says "this year".
16:32:18 <Ilari> Haha. Reading comment by Huston written in November: "A less conservative model that uses settings that reflect continued escalation of demand through 2011 now forecasts APNIC exhausting its address pools in September 2011."
16:33:26 <Ilari> Heck, now it is mostly question of if depletion occurs in first or second half of April (The current huston estimate of May 5th looks quite overly optimistic).
16:37:37 -!- cheater99 has joined.
16:38:36 -!- cheater00 has quit (Ping timeout: 248 seconds).
16:40:04 -!- cal153 has quit (Ping timeout: 260 seconds).
16:41:50 -!- sftp has joined.
17:01:28 <oklopol> "<cheater00> oklopol: wait, you're finnish??" <<< yes, what else?
17:02:56 <oklopol> i figured my decent knowledge of the finnish language was a dead giveaway
17:05:57 <olsner> everyone's finnish in here
17:09:44 <cheater99> oklopol: i know a little about suomi without being finnish!
17:10:35 <oklopol> do you know the language or the country?
17:11:43 <oklopol> i don't get why anyone would learn the language, but everyone seems to know "yksi, kaksi, kulma, nljy"
17:12:30 <Phantom_Hoover> I know that the country completely failed at mocking gullible foreigners.
17:13:51 <Phantom_Hoover> That's *classic* gullible foreigner material right there.
17:14:45 <Phantom_Hoover> I'm saying you shouldn't *unless* it enables you to mock gullible foreigners.
17:15:38 -!- asiekierka has quit (Ping timeout: 260 seconds).
17:15:40 <oklopol> maybe we could mock them by telling them we don't roll in the snow naked and when they go "figured" we could laugh behind their backs and go back to the sauna
17:16:01 <Zwaarddijk> what I don't get is swimming in holes in the ice.
17:16:25 -!- Mannerisky has quit (Ping timeout: 250 seconds).
17:16:49 <Zwaarddijk> but bagpipes were pretty common all over europe
17:17:03 <oklopol> Zwaarddijk: you've tried and disliked?
17:17:15 <Zwaarddijk> it's just the British Empire's Army's bagpipe corps that've made them associated with Scotland
17:17:23 <Zwaarddijk> oklopol: tried and liked and lapsed from liking
17:18:39 <oklopol> i haven't done it for a while either, although multiple times this winter
17:19:07 <cheater99> klop: i like the fact that suomi is similar to hungarian and japanese, from grammar
17:19:18 <oklopol> Phantom_Hoover: how? i still dgi
17:19:21 <Zwaarddijk> Phantom_Hoover: I figured that was the Scots tricking themselves into it
17:19:37 <cheater99> Phantom_Hoover: i've had really good haggis, but it was sold at waitrose in london, which is this posh supermarket.
17:19:46 <Zwaarddijk> sort of being gullible at being distinct from the english
17:19:48 <cheater99> Phantom_Hoover: i wouldn't dare eat it from say a stand or a pub somewhere.
17:20:53 <Zwaarddijk> cheater99: the similarity to japanese is less than commonly claimed
17:21:00 <oklopol> heat: i don't think we have anything in common with the japanese grammar, at least based on what i know sofar
17:21:17 <Zwaarddijk> oklopol: japanese almost has a case system
17:21:28 <cheater99> Zwaarddijk: knowing both japanese and finnish is less than common.
17:21:44 <Zwaarddijk> I know three our four people that know passable japanese and finnish
17:22:10 <Zwaarddijk> I've read typological accounts of Japanese
17:22:15 <oklopol> finnish is more head-last than english?
17:22:19 <Zwaarddijk> and I consider reading a reference grammar at some point
17:23:28 <oklopol> that indeed is a similarity
17:24:26 <Zwaarddijk> Phantom_Hoover: heads of phrases go after dependants
17:24:27 -!- cal153 has joined.
17:25:10 <Zwaarddijk> Phantom_Hoover: so like, English is head-first, since the head of a noun phrase (the article), goes first, the head of a prepositional phrase goes before the noun phrase, the head of the verb phrase (the verb) goes before the arguments
17:25:15 <oklopol> according to my tiny understanding, head-last = stack-based
17:25:27 -!- lament has joined.
17:25:40 <Zwaarddijk> however, both english and finnish are somewaht inconsistent
17:25:46 <Zwaarddijk> japanese is very consistently head-last
17:26:44 <oklopol> lament: that's a VERY cute smiley
17:30:02 -!- wth has joined.
17:30:06 -!- wth has quit (Remote host closed the connection).
17:30:07 <lament> you mean not head-last
17:30:22 -!- wth has joined.
17:30:34 -!- wth has left (?).
17:30:55 <oklopol> Phantom_Hoover: i think that was his joke to begin with
17:31:32 <Zwaarddijk> maybe i should try and get my hands on a japanese reference grammar
17:32:37 <oklopol> i should actually be studying japanese right now
17:32:48 <Zwaarddijk> I am part of the conlanging-community, and into typology and stuff like that
17:32:56 <Zwaarddijk> so it's not that far from my usual interests.
17:36:13 <oklopol> so what are the java and c++ of natural languages
17:36:36 <oklopol> and what's the ithkuil of esolangs
17:36:38 <Zwaarddijk> i would claim you can pick any two languages
17:36:53 <olsner> python might be Simple English
17:36:58 <Zwaarddijk> such that one of the is the java and the other is the c++
17:37:19 <oklopol> ithkuil is that thing where every feature of every existing language works in perfect unison
17:37:29 -!- wth has joined.
17:37:45 <oklopol> english sucks as much as the rest of them
17:37:55 <Zwaarddijk> Phantom_Hoover: I do, but I know several other langs as well
17:37:55 -!- wth has left (?).
17:38:03 <Zwaarddijk> and i find the most commonly cited reasons why English suck
17:38:10 <Zwaarddijk> are based on misunderstandings of how languages work
17:38:58 <Phantom_Hoover> Most programming languages suck, though, except those designed deliberately and carefully.
17:39:00 <oklopol> does that mean it's not the perfect language
17:39:22 <oklopol> natural languages suck all kinds of ass
17:39:34 <Zwaarddijk> natural languages are pretty well adapted to things though
17:39:44 <Zwaarddijk> I mean, look at the most "engineered" languages for human communication
17:39:55 -!- oerjan has joined.
17:40:34 <Zwaarddijk> but uh, there's a clear reason why we need redundancy, and that's why linguistic evolution so often converged on encoding things reduntantly
17:41:16 <Phantom_Hoover> Redundancy isn't so much my problem as incessant special-casing.
17:41:27 <Zwaarddijk> the special casing is a kind of optimization as well
17:41:34 <Zwaarddijk> it also does contribute to redundancy!
17:42:07 <Zwaarddijk> if all past tense verbs ended in -ed, that'd slightly increase the likelihood for mishearings
17:42:41 <Zwaarddijk> but uh, the special casing is often the result of a kind of inertia
17:43:13 <Phantom_Hoover> Yes, but that doesn't make memorising it any less stupid.
17:43:35 <Zwaarddijk> well, lots of verb forms that are regular are probably memorized anyway
17:44:28 <Zwaarddijk> there's some experiments that show that inflecting takes more time than recalling from memory
17:45:32 <Phantom_Hoover> I don't dispute that, but memorising special cases takes longer than memorising roots and inflection rules.
17:46:17 <Zwaarddijk> usually though, the irregular verbs are very commonly used ones
17:46:56 <Zwaarddijk> per verb, that is, not more often than the regular way of doing it
17:47:01 <Gregor> HEY VORPAL YOU SHOULD WRITE A VFS FOR MICROCOSM
17:47:20 <oerjan> <Gregor> Phantom_Hoover: Microcosm is a project that I refuse to make entirely my project, so it is as alive as other people are willing to let it be :P
17:47:33 <oerjan> so basically it's the ultimate experiment in lazy evaluation?
17:47:47 <Gregor> Phantom_Hoover: I ELECT YOU BUT ONLY IF YOU WRITE THE VFS
17:48:15 <Phantom_Hoover> NO VORPAL CAN WRITE IT AND I WILL MAKE SURE HE DOESN'T MAKE IT STUPID
17:48:47 <olsner> Phantom_Hoover: GOOD LUCK
17:49:12 <olsner> WE'RE LIKE THAT GUY IN DILBERT
17:50:23 <oklopol> I NEVER QUITE GOT WHAT THE POINT OF LOWERCASE IT
17:50:56 <oklopol> NOT REALLY, THIS WAY YOU DON'T HAVE TO PRESS SHIFT
17:51:08 <oklopol> WAIT, I NEVER DO THAT ANYWAY
17:52:12 <oklopol> I GOT THAT AFTER SAYING MY SAYINGS
17:52:32 <oklopol> SO YEAH I SUPPOSE THAT'S THE REASON
17:52:34 <Phantom_Hoover> THAT IS WHY THE ROMANS USED UPPERCASE: THEY WROTE BY CUTTING LINES INTO A MEDIUM
17:52:34 <oerjan> <Ilari> Haha. Reading comment by Huston written in November: "A less conservative model that uses settings that reflect continued escalation of demand through 2011 now forecasts APNIC exhausting its address pools in September 2011."
17:53:00 <oerjan> consider this a rehearsal of the singularity ;D
17:53:46 <oerjan> (if things turn asymptotic, i think very few human beings have any working intuition about it)
17:53:51 <oklopol> so i can't really parse Ilari's announcements, are there still ip4 addresses left?
17:54:13 <oerjan> oklopol: in the regional registrars, yes
17:54:29 -!- cheater99 has quit (Read error: Connection reset by peer).
17:54:54 -!- cheater00 has joined.
17:55:18 <oklopol> does africa have the same amount as europe?
17:56:04 <oklopol> because i think we have more internets than they do
17:56:10 <oerjan> europe is probably running out much faster, although they both got 1 /8 block at the global runout
17:56:30 <oerjan> (the europe RIR also includes the middle east)
17:57:12 <oerjan> africa is the smallest of them
17:57:22 <oklopol> does /8 mean a 256th of the whole space?
17:58:08 <oklopol> k i would've figured it'd be /24
17:58:13 <oerjan> APNIC (asian/pacific) is using much faster than the others again, though
17:58:30 <oklopol> actually maybe /8 is better
17:58:57 <oerjan> however as the last one to allocate normally, they got a bit extra at that point
18:00:38 <pikhq_> But they're still allocating at absurd rates.
18:00:50 <oerjan> <oklopol> i figured my decent knowledge of the finnish language was a dead giveaway
18:00:53 <Vorpal> <Gregor> HEY VORPAL YOU SHOULD WRITE A VFS FOR MICROCOSM <-- yes maybe during the summer, I don't have time now
18:01:06 <oerjan> well the rest us cannot _know_ it's decent, you could be just making gibberish
18:01:52 <Vorpal> oerjan, what about google translate? if it can't translate any of the words at all then it is likely to be gibberish for example
18:01:56 <oerjan> hakkapellittäan oklopoli on koskenkorvat
18:02:05 <oklopol> oerjan: i had a decent length conversation with Zwaarddijk in finnish tho. but i guess i could've planned this with him in pm
18:02:10 <Vorpal> oklopol, that looks plausible
18:02:56 <oerjan> that's due to my decent knowledge of the finnish gibberish
18:03:02 <pikhq_> oerjan ro'hìȳaku sannhiȳaku kuwasî ha ne! neko neko kawaî!
18:03:24 <Vorpal> so what does those words mean?
18:03:25 <oklopol> why would you put an there
18:03:59 <Phantom_Hoover> pikhq_, hey, you're not stupid; can you do the VFS for Microcosm?
18:04:02 <pikhq_> oklopol: darukadaruka muhame'tò sìha'tò no tè.
18:04:02 <oerjan> oklopol: well i thought there had to be a new root inside there, so the last part was frontal
18:04:07 <oklopol> i still have no idea what you meant though
18:04:13 <pikhq_> Phantom_Hoover: Can? Probably. Will? Probably not.
18:04:43 <Phantom_Hoover> pikhq_, dammit, that's two of us. Except if I did it all chance of Windows portability would be out of the window.
18:05:08 <Vorpal> <pikhq_> oklopol: darukadaruka muhame'tò sìha'tò no tè. <-- what is this?
18:05:38 <Vorpal> pikhq_, looks like a mix of Japanese and some language Tolkin made up :P
18:06:05 <oklopol> in the first one you almost said 600 300 then kuwa is some sort of farm related tool and then cat cat cute
18:06:06 <pikhq_> Vorpal: It's an attempt to transcribe Arabic-esque jibberish into Japanese that I fucked up because I need coffee.
18:06:31 <pikhq_> oklopol: "Oerjan 600 300 explain, right? Cat cat cute!"
18:06:44 <pikhq_> The last bit was, of course, because of non-Japanese otaku.
18:07:26 <pikhq_> ... Waaait, that's not it, is it.
18:07:30 <pikhq_> Fucking hell I need coffee.
18:08:25 <pikhq_> kuwasii "detailed, accurate, well-informed"...
18:09:27 <oklopol> sannhiȳaku <<< shouldn't you have the hi -> bi thing somewhere?
18:09:34 <Sgeo> Ok, I officially suck at determining that things suck
18:09:36 <oklopol> i still don't know this notation :\
18:09:40 <pikhq_> oklopol: I really need coffee.
18:09:42 <Sgeo> I had no opinion of Rebecca Black's "Friday"
18:09:49 <pikhq_> That should be sannhìȳaku, yes.
18:10:15 <pikhq_> That's how you encode moraic "n" in my notation.
18:10:34 <Sgeo> By not hating something that the world seems to hate.
18:11:06 <oklopol> pikhq_: that much i reverse-engineered
18:11:14 <Vorpal> Sgeo, what thing? Windows ME?
18:11:29 <oklopol> and actually the rest too, except for that n thing, i suppose
18:11:40 <pikhq_> Helps that it's quite regular.
18:12:11 <pikhq_> "Every kana is encoded in one or two characters". Easy.
18:12:13 <Sgeo> http://www.youtube.com/watch?v=CD2LRROpph0
18:12:22 <oklopol> kana means chicken in finnish
18:12:35 * Sgeo lols at all the critical comments being marked as spam
18:12:42 <Vorpal> Sgeo, might check later.
18:13:03 <pikhq_> It means temporary notation in Japanese. :P
18:13:14 <oklopol> actually katakana means use the chicken as a table in finnish
18:13:24 <oklopol> well. not exactly, but anyhow.
18:14:16 <lament> japanese people are too crazy
18:14:20 <Zwaarddijk> more like "set the chicken", where "set" is used as the verb in "set the table"
18:14:56 <Vorpal> Zwaarddijk, dining on a chicken? awesome idea
18:16:18 -!- Slereah has quit (Ping timeout: 246 seconds).
18:16:24 -!- asiekierka has joined.
18:18:42 <pikhq_> lament: No, English speakers are too crazy.
18:18:50 <pikhq_> The Japanese people just don't give a shit.
18:23:07 <Vorpal> Phantom_Hoover, not even Gene Ray?
18:23:15 <Vorpal> (a most unusual name too)
18:24:26 -!- Slereah has joined.
18:27:58 <Vorpal> Phantom_Hoover, yes but the combination sounds quite weird :P
18:29:58 <Gregor> "differently sane" X-D
18:39:14 <Gregor> Also, combining a common first name with a common last name does not necessarily yield a common name.
18:39:24 <Gregor> You don't see too many Nguyen McTavishes.
18:46:44 -!- Slereah has quit (Ping timeout: 255 seconds).
18:48:32 -!- Sgeo has quit (Ping timeout: 255 seconds).
18:52:44 -!- Slereah has joined.
18:55:58 <oerjan> sven smith is probably not that rare in scandinavia...
18:56:57 <oklopol> ooooooooooooooooooooooooooooooooooooooooooooooooooooooko
18:58:03 <oerjan> hm actually the top google hits seem to all be in english
18:59:07 <oerjan> svein smith on the other hand gives many norwegian hits (although that's with norwegian google)
18:59:45 <oerjan> hm actually that's probably mainly one person, svein smith-meyer
18:59:47 -!- ais523 has joined.
18:59:58 <oklopol> okokokokokokokokokokokokoko
19:00:05 <oklopol> okokokokokokokokokokokokokoko
19:00:27 <oerjan> who seems to be a ceo kind of guy
19:02:17 <Phantom_Hoover> There's a guy in my school who thinks the fact that itanimulla.com redirects to nsa.gov is incontrovertible proof that the Illuminati control the government.
19:02:56 <oklopol> because it's an anagram of illuminata?
19:03:36 <ais523> surely a redirect the other way would potentially be proof, but that way round is irrelevant?
19:04:57 <oerjan> um actually it's not an anagram, there's an i/a mismatch
19:05:03 <Phantom_Hoover> I tried to tell him that, but it's nigh impossible to reason people out of positions they didn't reason themselves into.
19:05:12 <oklopol> it certainly is an anagram of illuminata
19:05:47 <oerjan> like the illuminati, just more neutral
19:06:19 <oklopol> damnit, guy at #math stopped asking why the zero vector alone forms a vector space
19:07:06 <oerjan> you mean he actually got it?
19:07:08 <ais523> oklopol: it wouldn't form a very /useful/ vector space...
19:07:48 <oklopol> oerjan: no, "<weia> moses: do this exercise on your own (that's all there is to it): Let F be any field and X any singleton set. Show that there is exactly one way to endow X with a K-vector space structure (addition, K-scalar multiplication, zero)." "<moses> weia: i will" "<moses> bbl"
19:08:27 <oklopol> if you don't have the zero space, many things become more complicated to state
19:08:51 <oerjan> vector spaces are varieties, and all varieties have trivial members
19:09:19 <oklopol> also what oerjan said although that's essentially the same thing
19:09:39 <oklopol> variety = closed under subthings and products and homosexual images
19:09:57 <oklopol> so not having {0} would be a silly exception
19:10:14 <oerjan> that = should be an equivalence though, i don't think they usually take that as the definition
19:10:29 <oklopol> i saw it as a definition *today*
19:11:12 <oklopol> if you just have finite direct products
19:11:22 <oklopol> then you don't get that varieties = equation defined shit
19:11:29 <oklopol> that varieties are that in some eventual sense
19:11:52 <oklopol> i don't know what that sense is because the article about this couldn't be found on the internets :(
19:11:58 <oerjan> i don't recall but i may have touched by it
19:13:06 <oklopol> but were you touched by it?
19:13:09 <oerjan> i recall there's a connection between monads over Set and varieties with possibly infinite operations, though
19:13:11 <oklopol> i guess not if you don't recall it
19:13:57 <oerjan> i'm not sure if that could be similar to what you say
19:14:39 <oerjan> this i read in an encyclopedia (paper) article about monads, though
19:17:03 -!- copumpkin has joined.
19:18:01 <Ilari> APNIC relative allocations in last 30 days: Something like 86.8%. Insane. Just plain insane.
19:18:11 <oerjan> ok this is far too dense for my brain http://en.wikipedia.org/wiki/Monad_(category_theory)#Algebras_for_a_monad
19:18:18 <pikhq_> Some 0.25 /8s were allocated today, right?
19:18:41 <Ilari> That is, over 5 times the rate of entiere rest of the world combined.
19:18:53 <oerjan> i was hoping for something specialized to Set
19:19:07 -!- asiekierka has quit (Ping timeout: 255 seconds).
19:19:19 <Ilari> Last 30 days rate for APNIC is now something like 2.34 blocks.
19:19:51 -!- BeholdMyGlory has joined.
19:20:51 <Ilari> Which is something utterly insane.
19:21:40 -!- BeholdMyGlory has quit (Remote host closed the connection).
19:21:54 * pikhq_ builds GCC for reasons of hopefully awesome
19:22:40 -!- BeholdMyGlory has joined.
19:22:46 <Ilari> For this year as whole: 4.30 blocks.
19:25:58 <Ilari> 77 days. And over half of those addresses have been allocated in last 30 days. Wow.
19:26:05 <oklopol> oerjan: i got stuck at the adjoint functors part
19:26:10 <ais523> how many blocks does APNIC have left?
19:27:03 <Ilari> 1.84 (not couting setaside block).
19:27:23 <pikhq_> oklopol: I'll just assume that I will understand that much better when I get further in my studies.
19:27:57 <oerjan> oh hm every pair of adjoint functors gives a monad, and every monad arises from such a pair in at least one way
19:28:19 <oklopol> yeah but what the hell is a pair of adjoint functors
19:29:04 <oklopol> i'll continue trying to understand ->
19:29:07 <oerjan> well a typical example is the free functor and the underlying functor for a variety of algebras
19:29:19 * pikhq_ looks forward to having a generally-usable link-time optimising compiler
19:29:53 <oerjan> say L is the functor which takes a set to the free monoid on the set, and R is the functor which takes a monoid to its underlying set
19:30:46 <oklopol> what's the free monoid on a set, S goes to all finite sequences of elements of S?
19:30:50 <oerjan> then L : Set -> Monoid and R : Monoid -> Set is an adjoint pair, and RL : Set -> Set is essentially Haskell's list monad
19:31:12 <Ilari> At 2.34 blocks per 30 days, depleting that would take 24 days. 24 days from now is 11th April.
19:31:14 <pikhq_> I'm presuming if "GCC, Mozilla Firefox, and other large applications" work, then GCC's LTO is actually functional.
19:31:29 <oklopol> i guess no one said they'd be
19:31:31 <oerjan> no, they are not usually inverses
19:32:05 <pikhq_> elliott would be so happy; this makes static linking worthwhile. :P
19:32:07 <oerjan> there are natural transformations that make them almost inverses, though
19:32:16 <pikhq_> Heck, might even make static linking against *glibc* practical.
19:33:03 <oerjan> Id -> RL is one, and that's the return for the list monad
19:33:50 <oerjan> LR -> Id may be coreturn for the list comonad, i'm not entirely sure
19:34:11 <oerjan> (for the example above)
19:34:17 <oklopol> but what does it mean for a bijection to be natural?
19:34:27 <oklopol> trying to get the wp article and that's hard :<
19:35:11 -!- FireFly has joined.
19:36:13 <oklopol> mostly because i'm not looking up the definition of naturality
19:36:23 <oerjan> that it is a natural transformation, i presume
19:36:49 <oerjan> naturality is pretty fundamental for category theory
19:36:55 <oklopol> oh one of those things natural transformations have between hom sets?
19:38:09 <pikhq_> No *wonder* oerjan finds Haskell so natural. He actually knows category theory, rather than things vaguely related to it! :P
19:38:21 <oerjan> the bijection should _be_ a natural transformation when restricted to each variable (and taking the most obvious functors to be a transformation between)
19:38:55 <oklopol> a natural transformation between what functors?
19:39:53 <Phantom_Hoover> Someone give me a decent syntax for unordered pairs in Haskell.
19:40:08 <oerjan> lessee hom_C(FY, X) -> hom_C(Y, GX)
19:40:09 <oklopol> yeah idgi, i don't see two categories with the same "domain and codomain" categories
19:41:39 <oerjan> so if you fix X, say, the left side is the composition of the hom_C(., X) functor with the F functor and the right side is simply the hom_C(. , GX) functor
19:42:21 <oerjan> and the bijection being a natural transformation between those is what it means to be "natural in Y"
19:43:09 <oklopol> let's dwell in "the hom_C(., X) functor" for a while
19:43:43 <oklopol> that's a function from C's objects to certain subsets of its morphisms
19:44:12 <oklopol> how do you make it a functor?
19:44:50 -!- Sgeo has joined.
19:45:33 <oerjan> it's a functor from Set to Set, actually
19:46:13 <oerjan> which explains how this is all in one category
19:47:24 <copumpkin> we should legislate against unnatural transformations
19:47:40 <copumpkin> and I don't think they should be allowed to exist
19:47:47 * oerjan tries to make his memory work
19:48:28 <oerjan> or wait it's a functor from C to Set
19:49:17 <oerjan> an object Y in C is taken to the set hom_C(Y, X)
19:49:21 <oklopol> so you map an object Z to the set of morphisms from Z to X or what?
19:50:34 <oklopol> then if you have Z and Y in C, and morphism f : Z -> Y, then it goes to the set of morphisms you get by taking morphisms from Z to Y and adding a morphism from Y to X?
19:51:13 -!- Xuu has joined.
19:51:16 <oklopol> this covariance/contravariance thing is soooo complicated :D
19:51:18 -!- Zuu has quit (Quit: leaving).
19:51:31 <copumpkin> oklopol: not really, it's just functors on opposite categories
19:51:33 <oerjan> oh right one of those hom's is contravariant
19:51:41 <Phantom_Hoover> I'm trying to do unordered pairs, and I've just been told that (a ~ b) cannot be deduced.
19:51:58 <copumpkin> Phantom_Hoover: unordered heterogeneous pairs?
19:52:14 -!- Xuu has quit (Changing host).
19:52:14 -!- Xuu has joined.
19:52:14 <oerjan> the hom(., X) is contravariant iirc
19:52:28 -!- Xuu has changed nick to Zuu.
19:52:28 <copumpkin> oklopol: I've used that many times in the past :)
19:52:32 <oerjan> while hom(Y, .) is covariant
19:52:51 <oerjan> and yeah the functor on morphisms is just composition at the right side
19:53:53 <oklopol> but which one is contra and which is co
19:53:57 <ais523> oerjan: after all your and elliott's laughing about me about adding a computable-real type to C, I attended a seminar on Wednesday that demonstrated I wasn't talking nonsense after all
19:54:13 * oerjan doesn't recall laughing
19:54:45 <oerjan> oh you mean what it means
19:54:53 <oklopol> what does contra/co mean, those terms tell me exactly as much as Hom(., X) and the other
19:55:25 <oerjan> covariant is the "usual" functor, sending morphisms to morphisms in the "same" direction
19:55:42 <oerjan> contravariant switches the direction of the morphism on the end
19:56:13 -!- augur has quit (Remote host closed the connection).
19:56:38 <oerjan> so if you have a morphism f : A -> B, then F(f) : F(A) -> F(B) if it's covariant but F(f) : F(B) -> F(A) if it's contravariant
19:56:40 <ais523> the seminar even contained infinitary operations for a good reason
19:56:46 <oklopol> that's kinda obvious in the general case, but i can't seem to get it in this particular case with Set on the other end...
19:56:51 <ais523> normally, people just put infinitary operations in their languages to annoy my supervisor
19:57:48 <oklopol> that's just arbitrary of course
19:57:50 <oerjan> ok so if f : Y -> Z is a morphism in C, then since this is contravariant we want a morphism from Hom(Z, X) to Hom(Y, X)
19:58:18 <copumpkin> we have CT seminars at my job every friday :)
19:58:50 <oerjan> if g is in Hom(Z, X) then we use g . f
19:59:38 <oerjan> i recall we used the notation f^*(g) (in LaTeX)
20:00:27 <oerjan> and f_*(g) = f . g to make the other hom-functor Hom(Y, .)
20:00:45 <oerjan> (composing on the other end)
20:01:29 * oerjan learned all this originally for categories of modules over rings
20:03:59 <oklopol> okay i think i somewhat got what you said about naturality
20:04:06 <oklopol> but there's still some work to do
20:24:19 -!- augur has joined.
20:26:12 <oklopol> (i did other stuff too :P)
20:26:46 <oklopol> it's kinda hard to get all of these things in your head
20:27:03 <oklopol> you have two categories, and two functors, and Set and gah
20:27:13 <copumpkin> if you want category theory, I've implemented lots of fun stuff in agda
20:27:23 <copumpkin> in the most general way possible, unlike previous attempts
20:27:37 <copumpkin> for example, https://github.com/pumpkin/categories/blob/master/Category/NaturalTransformation.agda#L98
20:30:26 <oklopol> i'm sure this would be perfectly easy on paper, but i still find natural transformations kinda hard to picture when they occur in a "concrete" situation
20:30:49 <oklopol> the abstract definition is natural enough, so i'm sure i'd get it fast enough if i bumped into categories more often than ones every 2 months
20:30:57 <copumpkin> you should've been at our seminar today! people were having the same trouble there and I think we got rid of it
20:31:07 <augur> you and your seminars, copumpkin
20:31:40 <augur> wait, a you work with edwardk?
20:31:52 <oklopol> i have enough seminars to worry about as it is
20:32:21 <copumpkin> he's sitting across from me right now
20:33:48 <oklopol> so you could like, strangle him if you wanted to?
20:35:39 <oklopol> you should leave the room quickly and quietly, there are recorded cases of humans attacking other humans for no apparent reasons
20:35:49 -!- hagb4rd|afk has quit (Ping timeout: 248 seconds).
20:36:45 <copumpkin> oh, so I should just avoid all other humans?
20:37:44 <ais523> <oklopol> you should leave the room quickly and quietly, there are recorded cases of humans attacking other humans for no apparent reason <--- I like this quote, but HackEgo isn't here
20:39:04 <oklopol> copumpkin: i'm not saying you have to avoid humans, just keep an eye on them at all times
20:39:37 * augur makes copumpkin pie
20:40:00 <oklopol> and carry some money in your pocket so you can distract males if they get aggressive
20:40:47 <ais523> does money not work on females?
20:41:13 <oklopol> females are a bit trickier
20:41:30 -!- EgoBot has joined.
20:41:34 -!- HackEgo has joined.
20:42:02 * augur steals copumpkin's cookies
20:43:10 <augur> mein kampfy chair?
20:44:46 <augur> i aint chowin your kampf, gtfo
20:51:12 <ais523> (just in case my computer crashes again)
20:52:39 <ais523> hmm, that's Wed Apr 14 08:07:58 BST 2004
20:53:58 <lament> that was a great moment
20:54:57 <ais523> it lasted a whole second
21:09:49 <lament> i would read programming for philosophers
21:11:22 <Gregor> `echo ais523: I notice you don't even add the quote when he reappears :P
21:11:24 <HackEgo> ais523: I notice you don't even add the quote when he reappears :P
21:11:42 <ais523> `addquote <oklopol> you should leave the room quickly and quietly, there are recorded cases of humans attacking other humans for no apparent reason
21:11:43 <HackEgo> 336) <oklopol> you should leave the room quickly and quietly, there are recorded cases of humans attacking other humans for no apparent reason
21:41:11 <ais523> it's been relatively unfunny this year
22:30:25 <Phantom_Hoover> http://www.cnn.com/2003/SHOWBIZ/Music/03/14/dixie.chicks.reut/
22:42:45 -!- impomatic has quit (Quit: ChatZilla 0.9.86 [Firefox 3.5.17/20110121150729]).
22:54:33 -!- Sgeo has quit (Ping timeout: 255 seconds).
23:07:43 -!- copumpkin has quit (Quit: Computer has gone to sleep.).
23:15:42 -!- oerjan has quit (Quit: leaving).
23:17:36 -!- zzo38 has joined.
23:17:41 <zzo38> I did try out yoob.
23:30:49 <ais523> what did you think of it?
23:32:15 <zzo38> What I think is there should be a "load graph" for noit o' mnain worb. That is, there is a row on the graph for each ! and the pixel is lit when there is any ball in that position and dark otherwise.
23:34:40 -!- BeholdMyGlory has quit (Remote host closed the connection).
23:34:59 <zzo38> Also, it seems you cannot edit unless you load an example first.
23:49:26 -!- MigoMipo has quit (Read error: Connection reset by peer).
23:54:22 <zzo38> I do not want all the features of Wayland, and some things I would like changed. If I make a Linux system I could make the changes.
23:58:44 <zzo38> You said before that Ubuntu would change to Wayland. So I should make up a bit different kind.
23:59:04 <ais523> well, I doubt Ubuntu would be an ideal Linux distro for you anyway
23:59:29 <zzo38> Yes, I would make my own distribution instead of using Ubuntu or any other one.