00:12:46 <pikhq_> Gregor: Quick question: should I switch to aptosid, and if so how best to do so?
00:12:55 <pikhq_> Like, just edit sources.list and dist-upgrade?
00:13:10 <Gregor> I've never tried a switch-over to aptosid, but I can't imagine why it wouldn't work.
00:13:39 <Gregor> I like it. What are you switching from?
00:14:48 <Gregor> Yeah, I'd switch. I've been using Aptosid more-or-less hassle free since before it was Aptosid *shrugs*
00:17:40 <zzo38> Did they use the longitude/latitude numbers on Mars, to figure out how to land on Mars? I think they have defined where is zero longitude on Mars by now.
00:17:56 <pikhq_> Dammit Aptosid people, tell me what your sources.list magic is!
00:19:05 <pikhq_> I'm trying to do unconventional things here!
00:19:26 <Gregor> pikhq_: Dude, Download/Mirrors on the left.
00:19:42 <Gregor> It shows them all in terms of the exact lines to add to sources.list
00:23:21 <zzo38> How to cross-platform check for file exist and file size in a C code?
00:26:11 <pikhq_> "Need to get 623MB of archives." Wheeee!
00:27:28 <FreeFull> zzo38: I know with POSIX you're guaranteed to be able to do it, but I don't know if you can if one of your platforms isn't POSIX
00:29:27 <zzo38> FreeFull: This program ought to work on Windows and on UNIX (including Linux). This program also uses SDL, in case that matters at all.
00:34:36 <zzo38> (On Windows, it can be using MinGW and Cygwin, although Cygwin is also UNIX anyways.)
00:38:20 <zzo38> Where is document to explain which of these functions work in what system (MinGW, Cygwin, Linux, etc)?
00:43:04 <pikhq_> Gregor: Did I do this wrong, or is the latest version of Iceweasel in aptosid version 10?
00:43:49 <zzo38> Can stat() be used in MinGW and Cygwin and Linux and FreeBSD?
00:58:12 -!- itidus21 has quit (Read error: Connection reset by peer).
01:05:55 -!- pikhq has joined.
01:09:11 -!- pikhq_ has quit (Ping timeout: 244 seconds).
01:10:21 -!- pikhq has quit (Client Quit).
01:16:21 -!- pikhq has joined.
01:19:53 <pikhq> aptosid doesn't enable zram!
01:23:07 <oerjan> now with the nonsense words again
01:23:51 <pikhq> aptosid is a Debian variant, zram is a compressed RAM disk for use as swap.
01:24:25 <pikhq> Back to having a swap partition. :(
01:24:27 -!- Phantom_Hoover has joined.
01:24:27 <oerjan> my comment was really about how they sounded like nonsense :P
01:25:00 <pikhq> At least now I have a kernel sans retarded-broken writeback.
01:26:24 -!- Phantom__Hoover has quit (Ping timeout: 276 seconds).
01:26:27 <oerjan> which reminds me about this i saw linked in a reddit comment http://www.laurenipsum.org/mostly-lost
01:28:33 -!- derdon has quit (Remote host closed the connection).
01:37:27 <pikhq> *And* they didn't enable codel!
01:37:38 <pikhq> These people don't know how to configure a damned kernel.
01:44:42 <FreeFull> zzo38: stat() should work with Cygwin, Linux and FreeBSD fine. Not sure about MinGW
01:45:15 <zzo38> Well, I tried it just now with MinGW and it seems to be working OK.
01:52:27 <Gregor> pikhq: Aptosid doesn't have its own Iceweasel; it doesn't have most packages, just fixes. The latest version is whatever's in Debian sid (assuming you correctly switched testing->sid too)
02:33:51 <pikhq> Gregor: I was under the impression it might have upgraded some silly-old packages still.
02:34:56 <pikhq> Gregor: Ah, whatever, I don't use it as my primary browser anyways.
02:34:59 <Gregor> It's pretty hands-off.
02:35:16 <Gregor> *gives self points for not typing “hooves-off”*
02:41:53 * Phantom_Hoover sees someone on the internet arguing that Modern Warfare 3 is morally abhorrent because the button you press to kill a sleeping guard is the one most games use for miscellaneous interactions like flipping switches.
02:46:58 -!- Phantom_Hoover has quit (Quit: Leaving).
02:51:19 * oerjan notes the following reddit title "So excited for the showers tonight! :D (First picture I've ever taken)."
02:52:13 <oerjan> links to this http://imgur.com/lYxEr
02:56:01 <oerjan> it's understandable that they had to disable the downvote button in that subreddit.
02:58:52 <zzo38> Is it pretty hooves-off too or only hands-off?
03:01:08 <oerjan> it has nothing to do with ponies, anyhow
04:02:46 -!- oerjan has quit (Quit: Gah headache).
04:21:10 -!- amberdawn has joined.
04:21:26 -!- amberdawn has changed nick to Zelinc.
04:22:43 <zzo38> What is the best C code to convert the NES/Famicom attribute table address to the address of the tile which it affects?
04:24:06 -!- Zelinc has changed nick to amberdawn.
04:24:16 -!- amberdawn has left.
04:25:22 <zzo38> This is stupid why must the bit select operation be only for INTERCAL?
04:26:16 <zzo38> And also bit unselect operation which not even INTERCAL includes.
04:27:10 <pikhq> Huh, bit select seems like a generally useful operator.
04:27:44 <zzo38> pikhq: Yes, I agree.
04:28:04 <pikhq> I mean, it's slightly weird I guess, but that's actually something I could see myself using in real code.
04:28:21 <zzo38> Yes I would use it too.
04:29:37 <zzo38> But sometimes the reverse operation can be useful too
04:30:58 <zzo38> While writing Famitile I already found at least three uses of this bit select operator and its reverse.
04:34:09 <FreeFull> What does bit select do again?
04:35:08 <pikhq> FreeFull: n `select` mask is *kinda* like n | mask, except instead of the 0 bits in the mask being 0 in the result they are non-existent.
04:35:27 <pikhq> So, 101 `select` 100 = 1, instead of 100.
04:35:40 <FreeFull> So it's like an and and a right shift
04:36:28 <FreeFull> So would 1101 `select` 1011 be 101
04:37:08 <zzo38> What could be used here is the reverse bit select of 0x1DC, I think.
04:37:52 <zzo38> FreeFull: INTERCAL has that operation built-in. Other programming languages don't have that, and I know of no machine code having such instruction either.
04:37:53 <FreeFull> I really wish C had native rotate and reverse bits
04:38:36 <zzo38> FreeFull: Yes it can be useful to do so.
04:38:38 <FreeFull> zzo38: ARM has an instruction that takes an offset and number of bytes to extract, but that would only work for a continuous group of bytes
04:45:48 <FreeFull> There is also bit field insert
04:46:51 <FreeFull> Anyway, it wouldn't be too hard to write a function selectBits(int x, int y);
04:47:57 <zzo38> But it might be too slow
04:48:04 <FreeFull> Well, you'd probably want uint64_t selectBits(uint64_t x, uint64_t y);
04:48:11 <FreeFull> Wouldn't be any slower than intercal =P
04:48:57 <zzo38> Actually C-INTERCAL includes optimization files, so yes it would be slower than INTERCAL.
04:49:47 <FreeFull> Have you tested a well written selectBits() function against C-INTERCAL's `select` then?
04:56:46 <zzo38> No but it seems like it would be slower. Probably if the right operand is a constant it would be done better using other stuff. If the right operand is not constant then that is different.
04:57:13 <zzo38> But probably either way is slower than if there was a machine instruction for that purpose.
05:00:49 * Sgeo sads at the Bloom tutorial assuming I have a Bloom repl open
05:01:30 <FreeFull> The compiler would probably optimise it into some inline ands, ors and shifts
05:01:55 <FreeFull> I don't think x86 has an instruction that would correspond to select
05:02:41 <zzo38> Yes, they would do it like that, if it is constant.
05:04:05 <FreeFull> If not constant then probably some unrolled loop
05:04:06 <pikhq> FreeFull: x86 prefers to add instructions for crc32.
05:04:28 <pikhq> As well as memcpy.
05:04:43 <FreeFull> Well crc is cyclic redundancy check
05:04:51 <pikhq> memcpy can be *branchless* on x86.
05:06:46 <zzo38> Things like "'.1~.1'~#1" in INTERCAL might become (!!r1) in C, and "'.1~.1'~#2" indicates that the value of register 1 is neither zero nor a power of 2.
05:06:58 <pikhq> Though a more efficient implementation will use rep movsd or rep movsq to deal in 32-bit or 64-bit quantities at a time, and then copy any left-over bytes...
05:08:52 <FreeFull> Copying happens a lot in some contexts, so it made sense for the x86 people to add in the string instructions
05:10:12 <FreeFull> Why would someone have to do x: movsb loop x when rep movsb will do the same thing much quicker
05:11:02 <pikhq> Because that has actually no relation with how the CPU works at all.
05:12:05 <FreeFull> Well, if you're copying from memory to memory
05:12:15 <FreeFull> There isn't much reason for the data to go through the CPU
05:12:37 <pikhq> Where the hell else would it go? Memory isn't magic.
05:12:43 <pikhq> (though friendship is)
05:14:36 <FreeFull> Through a dedicated memory chip?
05:14:52 <pikhq> What, you mean a CPU?
05:16:44 <FreeFull> Well, I guess the northbridge is integrated into the CPU nowadays
05:17:03 <zzo38> Then they should make it faster by allowing you to put your own microcodes in and modify them at runtime, and including the instructions from esolangs Muxcomp and Checkout, as well.
05:17:06 <pikhq> It was also little more than a translator between the CPU bus and the RAM bus.
05:17:07 <FreeFull> It only got integrated in in 2011
05:17:22 <pikhq> Much earlier if you use AMD.
05:17:36 <FreeFull> My point is that a CPU is much faster than RAM
05:17:59 <FreeFull> So while RAM is being copied, it can do something else
05:18:20 <pikhq> So rep movsb is wasted silicon, because actually doing the operations that it's doing in microcode takes no more time?
05:18:26 <pikhq> Thank you, I agree. :P
05:20:15 <pikhq> Also, the CPU might be much faster than RAM, but the RAM is only operating in terms of "load $small_amount $address" and "write $small_amount $address".
05:20:40 <pikhq> ($small_amount on modern systems being 64 bits)
05:21:29 <pikhq> And what better to deal with the need to emit the fetches and writes to gradually incrementing addresses than the processor?
05:22:05 <pikhq> Especially keeping in mind that cache also comes into play here, and that's controlled entirely *in* the processor.
05:22:58 <FreeFull> Why would the data go into the cache when the processor isn't actually doing anything with it, just moving it around?
05:23:56 <pikhq> Even if *somehow* by *magic fairy dust* it can "not do anything" but "move it around", *cache still comes into play*. Cache coherency, have you heard of it?
05:24:52 <pikhq> Basically, the cache in the CPU must correspond to the actual state of RAM, otherwise you get incorrect behavior.
05:26:00 <pikhq> Having an external thing handling the memory means either the CPU's basically blocking on that external thing or it's constantly getting cache invalidations over the bus.
05:26:00 <FreeFull> So if cache is pointing at a piece of ram that you're copying to, you need to update, otherwise you can keep the cache as it is
05:26:54 <pikhq> And in any case any writes to either area of memory will have to block on it in order for semantics to be preserved.
05:27:22 <pikhq> This is a *lot* of silicon that would be wasted for a dubious optimization.
05:27:38 <pikhq> And silicon space is a lot of what a CPU design actually optimizes for.
05:28:21 <FreeFull> What, not erasing cache when the cache isn't pointing at RAM that's being modified?
05:29:09 <pikhq> Blocking on the memory copy when you try writing over its source or destination before the memory copy is complete.
05:29:26 <pikhq> Merely only invalidating certain cache lines isn't *that* hard.
05:30:20 <zzo38> But what happen if it is RAM being modified by other devices or it is not RAM but is connected to something else instead?
05:30:21 -!- Phantom_Hoover has joined.
05:30:51 <pikhq> zzo38: Cache invalidation happens, and if you happen to be modifying RAM being modified by other devices your code breaks.
05:31:04 <pikhq> zzo38: The issue is, the second bit *is impermissible for memcpy()*.
05:31:35 <pikhq> FreeFull: Anyways, this is Premature Optimization.
05:32:12 <pikhq> FreeFull: If you show me a program where fast memcpy would make it much faster, I'll show you a program authored by a moron.
05:32:27 <zzo38> I mean the memory might be connected to I/O and ROM and timers and stuff instead of only being RAM.
05:33:59 <FreeFull> pikhq: In olden times, you'd want your memcpy to be as fast as you could go for graphics purposes, but I imagine nowadays graphics cards have their own inbuilt internal memory routines, and it doesn't matter as much
05:34:58 <FreeFull> zzo38: I don't think the CPU caches non-RAM though, does it?
05:35:11 <pikhq> In olden times, if you used memcpy for your graphics routines you were stupid.
05:35:46 <FreeFull> pikhq: So how are you supposed to display the same thing in two places on the screen then
05:36:22 <zzo38> I don't know much about CPU cache. But I suppose yes it could be designed to cache ROM, as long as it is not bank-switched, I guess.
05:36:47 <FreeFull> Well, if you're reading a lot from ROM for some reason, it makes sense to cache it
05:36:53 <pikhq> Pssst, memcpy doesn't help that much with that, and if you need to display something other than a rectangle in two different places *you can't use memcpy anyways*.
05:38:13 <pikhq> Yes, and congrats, you're using memcpy with fixed small size meaning it's very fast and probably doing it on memory that's cache which means it's fast anyways.
05:38:18 <pikhq> And you're still copying a rectangle.
05:38:55 <pikhq> It is very rare that you want to display a rectangle at multiple points in the screen.
05:40:37 <pikhq> And it's still premature optimization. Also, your utterly psychotic plan to have external hardware do the memcpy instead would only even be *faster* than the CPU just doing it if the amount to copy is large and/or not in cache anyways...
05:40:40 <FreeFull> Although now the graphics card would probably do a lot more of the heavy-lifting
05:41:02 <pikhq> If it's small and in cache, the CPU is also the cheapest thing to do it. Like, a dozen clock cycles or so.
05:41:20 <FreeFull> pikhq: I failed to consider that you'd need to keep the cache coherent
05:41:45 <pikhq> Which is almost enough time to emit a single fetch to RAM.
05:43:08 <FreeFull> Separate hardware would probably only make sense with a CPU that has no cache, only registers
05:43:45 <pikhq> And is designed to maximize performance, rather than cost/performance ratio.
05:44:08 <Phantom_Hoover> <pikhq> It is very rare that you want to display a rectangle at multiple points in the screen.
05:44:30 <pikhq> Phantom_Hoover: Sprite drawing is what you want to be fast.
05:45:03 <pikhq> At least, on old stuff like that.
05:46:30 <pikhq> Well, yeah, just drawing a backdrop is done once, unless you're scrolling the screen or something. ... In which case you actually want your graphics hardware to be able to do that for you on a slow system.
05:47:52 <FreeFull> I thought those old graphics cards only support vertical scrolling
05:48:48 <pikhq> Which is why you didn't see scrolling platformers on those systems.
05:49:01 <pikhq> The NES, for instance, did both.
05:49:32 <FreeFull> Well, the NES was made for games
05:49:46 <pikhq> Y'know, the thing where 2D graphics performance matters.
05:50:28 <FreeFull> IBM PCs ended up with games rather coincidentally, so initally didn't have specialised hardware for game stuff
05:51:11 <pikhq> They also were profoundly limited by the platform until CPU speed went high enough that a framebuffer was actually reasonable for it.
05:51:57 <pikhq> (the NES *did not actually expose a framebuffer*, and neither did other consoles of the day)
05:52:10 <FreeFull> Well, there was Mode X, which allowed you to touch four pixels at once
05:53:51 <zzo38> But it may be possible to make up a NES mapper hardware for exposing a framebuffer. MMC5 does a few things with the PPU nametables so possibly can be done other things, too.
05:54:24 <FreeFull> You can look at whatever NES demosceners do
05:55:01 <pikhq> Lots of clever tricks with the sprite and background tiles.
05:55:23 <pikhq> And, of course, changing the palletes.
05:55:39 <zzo38> Yes they do those things too.
06:25:01 <fizzie> FreeFull: You can do "hardware-accelerated" horizontal scrolling also with VGA, it's just slightly trickier. (Basically, twiddle the row pith register to get some invisible working room, increment/decrement the starting address by one to scroll right/left by 4 pixels, and use the Horizontal Pixel Panning register to scroll by 0..3 pixels to make it smooth.)
06:27:07 <fizzie> (As opposed to vertical scrolling which is just a matter of changing the starting offset.)
06:29:06 -!- nooga has joined.
06:31:17 <pikhq> fizzie: Oh, hey, that actually does work. Barely.
06:34:17 <fizzie> And what's this about a "psychotic plan" to have external hardware do a memcpy so that the CPU doesn't need to bother; isn't that pretty much exactly what DMA controllers are? (Okay, so they are mostly involved in I/O contexts, but you can do memory-to-memory DMA too. E.g. Intel has the http://en.wikipedia.org/wiki/I/O_Acceleration_Technology thing on their mobos for offloading memory copying.)
06:36:43 <pikhq> fizzie: He was suggesting it as an implementation of rep movsb.
06:37:04 <pikhq> fizzie: Y'know, the memcpy opcode.
06:37:43 <pikhq> You could do it, but there is no way you would want it for memcpy, and it could not sanely have the expected semantics of memcpy.
06:38:16 <fizzie> Well, okay, that'd be kind of weird.
06:38:34 <pikhq> We weren't talking the concept of memcpy, but literally void *memcpy(void *dest, const void *src, size_t n);
06:40:37 <fizzie> I had a look at the glibc x86-specific memcpy a while ago. It's amusingly complicated.
06:41:41 <pikhq> I bet it's also amusingly inefficient.
06:42:02 <Phantom_Hoover> Does it, like, special-case memory based on the first byte of the address because of some esoteric, tiny hardware effect?
06:42:02 <lambdabot> Phantom_Hoover: You have 1 new message. '/msg lambdabot @messages' to read it.
06:42:23 <pikhq> glibc is practically based on premature optimization and pessimization.
06:47:15 <zzo38> Actually I am currently writing a program for NES graphics, it is almost finish first version.
06:48:31 -!- zzo38 has quit (Remote host closed the connection).
06:48:49 <fizzie> Not the first byte, but I guess you could say it special-cases on the last byte since SSE-based copying is oh-so-fast when dealing with 16-byte aligned stuffs.
06:49:02 <fizzie> And the GCC builtin memcpy is not much less complicated.
06:50:13 <pikhq> fizzie: It's probably faster to just rep movsd or rep movsq. Certainly going to be quicker if memcpy is only occasionally called, in which case the impact on cache from the code matters...
06:52:52 <fizzie> I'm pretty sure they *do* do benchmarks with approximates-real-world-code uses, not just guess.
06:52:55 <fizzie> The GCC folks, at least.
06:54:53 <pikhq> The GCC folks perhaps. The glibc folks have had "performance improvement" commits that don't even function correctly.
06:55:10 <pikhq> Much less work faster.
06:57:11 <olsner> IIRC, the performance of rep movs compared to fancy memcpys have varied between "just as fast" and "ridiculously slow", depending on the cpu model
06:57:19 <olsner> it is still short though
06:58:49 <pikhq> And that matters heavily on modern systems.
06:59:12 <pikhq> If the disparity between RAM and CPU speed grows much more we might start switching to threaded code just to use less RAM.
07:04:57 <olsner> if the compiler just had any clue about which code to bother with, it would be free to do that for all the code that isn't run often
07:05:36 <olsner> a modern cpu is basically a jit compiler though, so maybe it could just do that all the time
07:06:21 <pikhq> The CPU would still have to fetch the code.
07:06:27 <pikhq> *Which is the slow part*.
07:11:00 <olsner> hmm, is all non-size optimization already pessimisation then?
07:11:58 <fizzie> I highly doubt you could have a rule like that.
07:12:58 -!- asiekierka has joined.
07:14:41 <fizzie> I haven't e.g. heard of ICC giving up loop unrolling completely; if only code size mattered, that'd be an obvious pessimization.
07:15:20 <fizzie> ICC's -O1: "Optimize to favor code size and code locality. Disables loop unrolling. -O1 may improve performance for applications with very large code size, many branches, and execution time not dominated by code within loops. In most cases, -O2 is recommended over -O1."
07:15:39 <fizzie> (-O2 is still the default.)
07:16:22 <asiekierka> at least it has a beta System Software (one that's not available anyone else)
07:18:29 <olsner> loop unrolling could expose e.g. vectorization or merge several loop iterations into one operation, and the compiler can undo it after it notices nothing happened except code getting duplicated (not sure if it does though)
07:18:32 <pikhq> fizzie: It's somewhat more complex than *always* being a pessimization, it's just that it *often* is...
07:19:18 <pikhq> fizzie: And if your goal is good software, not just FAAAST -funroll-all-loops!!!, then it's a pessimization for other factors your care about.
07:19:50 <pikhq> Like "being able to come back to it three months later and change something".
07:20:32 <fizzie> Uh.. why would loop unrolling by a compiler affect the maintainability of the source code?
07:20:46 <pikhq> fizzie: Sorry, I wasn't referring to loop unrolling in specific, but code size in general.
07:20:54 <pikhq> fizzie: Which the human writing source code has impact on.
07:21:10 <pikhq> My mind went off on a slightly different context than, uh, conversation would suggest.
07:21:21 <olsner> and I think the performance hit for fetching instructions is per cache line, so if you replace 1 byte of code with <32 bytes of faster code it might still pay off
07:22:10 * olsner does some handwaving and says something about caches
07:22:26 <pikhq> olsner: True, a cache line is effectively the minimum level of granularity for this purpose.
07:23:50 <olsner> assuming the loop runs for long enough that it ends up worth evicting some other code from the cache and loading it from RAM later
07:24:50 <pikhq> olsner: Simply loading it evicts something else, doesn't it?
07:26:06 <olsner> yes, and something always needs to be loaded... but if the code was "1 byte", there'd be 31 bytes left in that cache line that can be used for other things, and some reduction in eviction and loading
07:29:01 <olsner> I wonder how you could measure the amount of instruction-cache misses that are caused by bloated optimized code
07:30:09 <fizzie> By measuring cache misses with and without various optimizations, and comparing the numbers?
07:31:23 <olsner> yeah, but without recompiling everything :)
07:33:00 <fizzie> That's not much of a chore if you're serious about benchmarking. (Anyway, you can measure by Googling for someone who has already done it.)
07:36:26 <fizzie> (Sadly, first I hit found which compared miss rates with different GCC optimization levels was from 1995, that's not so very current.)
07:36:37 <fizzie> (I'll have to get going, I think.)
07:36:37 <kmc> i prefer to answer all performance questions according to a 1970's model of compilers and computer architecture, with zero empirical data
07:37:37 <olsner> I prefer to use plain old guessing
07:38:08 <fizzie> olsner: I prefer to call that the Stetson-Harrison method.
07:38:34 <olsner> I'm guessing that's a fancy name for guessing then
07:38:55 <kmc> also good fun is complaining about which C code will produce "more instructions" in a context where performance is completely irrelevant
07:38:59 <kmc> and also without reading the disassembly
07:39:32 <fizzie> olsner: It's a name for "pulling numbers out of a hat", really.
07:39:46 <fizzie> Not related, but see editor's note and coauthor list: http://www.improbable.com/airchives/classical/articles/peanut_butter_rotation.html
07:39:51 <olsner> sometimes you will find that introducing actual information into a discussion will quickly make it boring
07:41:14 <kmc> it's still a stretch to say that modern CPUs are JIT compilers, though
07:41:17 <kmc> as far as i know
07:41:35 <kmc> JITs will compile an entire function or a trace at once, and apply nonlocal optimizations within that
07:41:42 <olsner> yeah, it's a bit of an exaggeration
07:41:47 <kmc> CPUs only look ahead some number of instructions
07:41:57 <kmc> it's a reasonably fair exaggeration though
07:42:02 <kmc> and counteracts some pervasive myths
07:42:39 <kmc> man i kind of like programming on AVR where it actually just executes each instruction in order at a deterministic rate with no cache or anything
07:42:47 <olsner> I guess it's more like an on-the-fly translation with a cache, peephole optimizations and some reordering of instructions
07:43:56 <Phantom_Hoover> Was there some sort of new cookie legislation passed lately?
07:44:03 <kmc> in the UK yeah
07:44:17 <kmc> sites have to get your permission to set cookies
07:44:20 <fizzie> Phantom_Hoover: Yeah, chocolate-chips now require a registration.
07:44:26 <kmc> cause it's not like web browsers already let you control who can set cookies
07:44:32 <olsner> cookie legislation made me think about that cake/biscuit lawsuit
07:44:33 <fizzie> (It's like owning a gun.)
07:45:33 <olsner> this one claims >15.3% of time spent waiting for icache fills: http://ha.redhat.com/support/wpapers/redhat/OProfile/oprofile.pdf
07:46:17 <olsner> that seems to be from 2002, so it was basically produced last year
07:47:35 -!- AnotherTest has joined.
07:47:58 <fizzie> Ten years, one year: no difference.
07:48:05 -!- AnotherTest has left.
07:48:07 -!- AnotherTest has joined.
07:48:12 <fizzie> It's comparing GCC 2.95 and 3.1, that's very current indeed.
07:48:18 <fizzie> (What are they at now, 4.7?)
07:48:30 <olsner> The Matrix was released a year ago in 1998, so last year is at least 14 years long
07:48:48 <fizzie> olsner: Also doesn't compare between optimilization levels. :/
07:57:22 <shachaf> I'd forgotten how nice it was.
07:57:41 <kmc> it's all right, I guess
07:57:54 <kmc> maybe you should tell me why it's nice and I will hack my next trip to berkeley using this information
07:58:49 <shachaf> Maybe I just have low standards coming from (E)PA.
07:59:06 <shachaf> All I did was see part of a play and walk around.
07:59:36 <fizzie> Some people suggested I should spend 6-9 months at http://www.icsi.berkeley.edu/icsi/ but I skippeded it. :\ :/
08:00:14 <fizzie> (They have some sort of a collabomaration project at the university.)
08:03:28 <kmc> where did you walk around?
08:03:43 <kmc> east bay is all right but it reminds me too much of los angeles
08:03:49 <kmc> also athens reminded me of los angeles
08:04:07 <kmc> perhaps my principal component for cities is "how much is it like los angeles"
08:04:16 <shachaf> I was in Los Angeles a long time ago.
08:04:57 <shachaf> Maybe I should go there sometime.
08:05:13 <kmc> maybe you should go to helsinki instead
08:05:21 <kmc> preferably tomorrow or wednesday
08:05:42 <shachaf> I could be in Los Angeles tomorrow pretty easily if I wanted to.
08:05:55 <kmc> one day you'll be able to take Supertrain to los angeles
08:05:55 <shachaf> I could probably also be in Helsinki within 24 hours if I really really wanted to.
08:06:13 <kmc> "as your attorney i advise you to rent a very fast car with no top"
08:07:17 <kmc> one day you'll be able to smoke pot and get gay married on supertrain
08:07:21 <kmc> but no happy meals
08:07:29 <fizzie> One day you'll be able to take a Supertrain to Helsinki.
08:07:40 <fizzie> A supersonic, flying Supertrain.
08:07:43 <kmc> from tallinn? i hope so!!
08:07:53 <kmc> i proposed that railway here yesterday
08:08:18 <shachaf> Oh, you were proposing a superrailway? I thought it was just a regular railway.
08:08:24 <shachaf> Maybe supertrains go on regular railways.
08:08:45 <kmc> it would be a Train à Grande Vitesse
08:08:46 * shachaf recalls that kmc cares more about train infrastructure than about trains themselves.
08:08:52 <kmc> that's accurate
08:09:12 <coppro> kmc: you should come to canada and fix our rail system
08:09:15 <kmc> supertrain of the short lived awful 1979 tv show did not run on regular railways
08:09:16 <fizzie> The... uh, if the channel-tunnel is the chunnel... the HellTunnel? Yeah, that sounds about right. So the helltunnel's been proposed a few times in a reasonably serious manner, also somewhat recently. (Not so seriously that they'd actually get around building it, though.)
08:09:39 <kmc> but supertrain my idiosyncratic name for california high speed rail will run on pretty standard high speed rail lines
08:09:55 <fizzie> They went as far as to calculate some prices, came up with 1.1-1.6 billion euros in 1997.
08:09:57 <kmc> which is to say overhead electrified standard gauge track, grade separated and with wide curves
08:10:03 <shachaf> kmc: By that day it seems pretty likely that I won't be living in the bay area anymore.
08:11:17 <shachaf> kmc: If I go to Helsinki by Wednesday maybe I could mean KeithW.
08:11:46 <shachaf> Probably going to Boston would be an easier way to do that.
08:12:12 * shachaf doesn't have anything in particular to say to KeithW, though.
08:12:38 <kmc> coppro: is that more feasible than fixing the us rail system
08:12:40 <fizzie> I think there's a "fast train" kind of route being planned/under construction/something like that from Tallinn to Warsaw these days; that has sort of re-aroused interest in the Tallinn tunnel.
08:13:27 <fizzie> "Rail Baltica is one of the priority projects of the European Union: Trans-European Transport Networks (TEN-T). The project is supposed to link Finland, the Baltic States and Poland and also improve the connection between Central and Eastern Europe and Germany. It envisages a continuous rail link from Tallinn (Estonia), to Warsaw (Poland), going via Riga (Latvia) and Kaunas (Lithuania). It ...
08:13:33 <fizzie> ... will by-pass the Kaliningrad Oblast (Russia) and Hrodno (Belarus) where the two historic rail routes Poland-Lithuania have been going."
08:13:50 <fizzie> (It'll just be ferries for the Helsinki-Tallinn part so far, but maybe one day!)
08:13:54 <kmc> it might be freight-only :/
08:14:44 <kmc> wow, baltic countries don't even use standard gauge
08:14:45 <shachaf> But kmc prefers railways to water.
08:14:49 <kmc> they use russian gauge not surprisingly
08:14:54 <kmc> i like ferries too
08:15:18 <kmc> this trip has involved a fair number of ferries
08:15:39 <shachaf> WA has a bunch of ferries.
08:16:08 <fizzie> We had a Helsinki-Stockholm-Copenhagen-Amsterdam-and-so-on trip few summers ago, and it does take quite a long time to get the Helsinki-Stockholm-Copenhagen part done, mostly due to the slow boats.
08:16:35 <fizzie> Stockholm-Copenhagen trains were those very futuristically named "X2000" ones.
08:16:57 <kmc> shachaf: I took a train on a boat once!
08:17:21 <fizzie> (At least in 1990 "2000" was presumably futuristic.)
08:17:40 <kmc> København to Berlin train
08:18:01 <shachaf> Something like http://upload.wikimedia.org/wikipedia/commons/4/43/Woodf3a.jpg ?
08:18:10 <shachaf> kmc: Did you read _Three Men in a Boat_?
08:18:21 <shachaf> You should read that book!
08:18:33 <kmc> that's a nice picture
08:18:53 <fizzie> VR (the Finnish railway company) has deals called "BoatTrain", but I think it only means a discount for the train trip from wherever you happen to be to the harbour where the ships leave from.
08:19:09 <shachaf> It's from 1889 so you can get the text for free, too.
08:19:58 <kmc> i got one of those deals for my Dublin → Slough trip
08:20:06 <kmc> shockingly good deal
08:20:16 <kmc> it was €52 for the ferry and british rail ticket together
08:20:24 <kmc> compared to like £70 for the latter alone
08:20:34 <shachaf> http://www.gutenberg.org/ebooks/308
08:20:58 <shachaf> I think it's famous in Russia for some reason.
08:21:29 <fizzie> Our BoatTrains are just 50% off for the train ticket if you have a (regularly purchased) boat ticket.
08:22:52 <kmc> gutenberg.org gives a 403 if i use wget
08:23:14 <kmc> happily they serve a zip of the html version
08:23:20 <kmc> but i still can't wget it >_<
08:23:27 -!- AnotherTest has quit (Quit: Leaving.).
08:23:54 <fizzie> That's funny, I think I wget'd a whole lot of Gutenberg (maybe a mirror?) back when doing some text analysis stuff.
08:24:07 -!- AnotherTest has joined.
08:24:08 <kmc> well, did you lie about the user agent? ;)
08:24:46 <fizzie> http://www.gutenberg.org/wiki/Gutenberg:Information_About_Robot_Access_to_our_Pages
08:25:12 <fizzie> Come to think of it, I think I used the rsync method described in the mirroring page.
08:25:39 <shachaf> Their robots.txt does say crawl-delay: 2
08:25:57 <fizzie> Anywy, I'm sure you can wget from the ftp://ftp.ibiblio.org/ URLs. (Don't know if they generate any links there though.)
08:26:06 * shachaf is becoming increasingly frustrated.
08:26:18 <kmc> frustrated with what?
08:26:51 -!- epicmonkey has joined.
08:31:20 -!- liljuska has joined.
08:34:10 <kmc> i feel like that sometimes
08:44:49 -!- AnotherTest has quit (Ping timeout: 265 seconds).
08:48:17 -!- derdon has joined.
08:49:15 -!- Taneb has joined.
08:54:39 -!- itidus21 has joined.
08:59:23 -!- Taneb has quit (Ping timeout: 272 seconds).
09:01:03 -!- Taneb has joined.
09:03:28 -!- Taneb has quit (Client Quit).
09:03:29 <kmc> shachaf: do you want my half-baked advice?
09:10:04 <itidus21> i rode a train today. i had hot chips and a coffee, and i positioned myself to eavesdrop on an old man talking to a railway employee about a scheduled train not running due to vandalism
09:10:17 <kmc> traveling and change of scenery
09:10:24 <kmc> are both good for not being generally frustrated
09:10:34 <kmc> well, in particular if you have no clear reason to be frustrated
09:10:44 <kmc> if you're frustrated because of some significant problem in your life, that's different
09:11:07 <shachaf> That's probably good advice.
09:11:20 <kmc> it's probably eventually correct
09:11:47 <kmc> i also recommend psychedelic drugs for more or less the same reason
09:11:52 <kmc> change of internal scenery, or something
09:11:58 <kmc> not so much the visual, but metaphorically
09:12:21 <kmc> itidus21: Hot Chip is a good band
09:12:24 <kmc> itidus21: which train was not running?
09:12:52 <itidus21> 3:52pm train from melbourne flinders street to pakenham
09:12:57 <kmc> shachaf: do you have a job these days? do you like your job?
09:13:09 <kmc> australia eh
09:13:19 <kmc> it's, like, winter there
09:13:22 <kmc> how's that going
09:13:58 <itidus21> my psychologist encouraged me to get a jacket drycleaned
09:14:38 <itidus21> yeah, therapy is pretty cool if it's free
09:14:48 <Phantom_Hoover> itidus21, dry cleaning salesmen make poor psychologists.
09:15:36 <kmc> are you looking for a job?
09:15:42 <itidus21> well i was walking along in this jacket the other day and some woman in an SUV said get a job you bum
09:15:51 <kmc> i like xkcd today
09:15:54 <kmc> haha itidus21
09:16:18 <itidus21> its amazing the adventures that happen when you go out
09:17:08 <shachaf> kmc: Not actively but maybe I will soon.
09:17:43 <kmc> a friend of mine just quit his job
09:17:46 <itidus21> there was this guy who was begging for change with an elaborate story spoken quickly. he got $2 from me, and i saw two other guys hand him some kind of money eventually
09:17:46 <kmc> as in today is his last day
09:18:21 <kmc> he says the younger employees understand why he's doing it
09:18:28 <kmc> and the older employees think he's totally crazy for quitting without another job lined up
09:18:40 <Phantom_Hoover> itidus21, one of Jon Ronson's old Guardian columns has him being groomed by a conman for 2 hours at an airport for $60.
09:18:46 <kmc> my friend is young and doesn't have kids and has savings and is a rails developer in SF
09:18:49 <kmc> so i think he will do ok
09:19:07 <kmc> $30 an hour ain't bad!
09:19:18 <kmc> wonder what the actual rate is, with downtime
09:19:58 <itidus21> in a rapid delivery he explained without pausing for a breath that he's not after cigarettes, that he needs $1.40 for a bus, something about having trouble with a flatmate, that if he sees me sometime he'll give me a few $
09:21:03 <itidus21> then he walked off.. and i saw him approach 2 people before going inside the shopping center, then later coming out and approaching another and going back in
09:22:20 <itidus21> shopping centre is what they call a mall here i think
09:23:52 <shachaf> Maybe I should do the university thing.
09:24:03 <shachaf> There are complications involved, of course.
09:25:32 <itidus21> there was also this girl busking with hoop dancing
09:25:42 <kmc> shachaf: do you have a bachelor's degree already?
09:26:14 <itidus21> i guess. she wasn't very talented either. but points for trying
09:26:31 <kmc> hoop dancing is cool
09:26:43 <kmc> i heard that busking licenses are graded by instrument
09:26:48 <kmc> and that bagpipe license is the most expensive
09:27:11 <kmc> in New York there are a lot of guys who breakdance on the floor of moving subway trains
09:27:13 <itidus21> Phantom_Hoover: is that for the tourists?
09:27:18 <kmc> occasionally a rider gets kicked in the face
09:27:22 <itidus21> yeah... theres guys here busking with didgeridoos
09:27:29 <kmc> in Boston they have this too except that it's just one particular guy who is everywhere
09:27:39 <Phantom_Hoover> itidus21, presumably, because you'll never find anyone doing it except on Princes Street.
09:27:55 <kmc> there's a guy who plays bagpipes in FiDi in SF
09:27:59 <kmc> i don't know if he's busking or what
09:28:59 <itidus21> there was also a guy with an accordian if i recall
09:29:17 <itidus21> thats part of the fun of going to the city
09:29:25 <kmc> in new york you also see people busking with some kind of traditional chinese string instrument
09:29:28 <kmc> which sounds awful
09:30:00 <shachaf> Is it the two-string thing?
09:30:08 <kmc> and people doing traditional chinese singing in the park, which also sounds awful
09:30:09 <Phantom_Hoover> So my image of the cities of Australia is a load of empty streets crammed with buskers and conmen.
09:30:12 <kmc> shachaf: maybe?
09:30:13 * shachaf tries to recall what that's called.
09:30:16 <kmc> it's like an upright thing with a bow
09:30:25 <itidus21> Phantom_Hoover: ahh well the conman was in a suburb :P
09:30:45 <shachaf> https://en.wikipedia.org/wiki/Erhu
09:30:51 <itidus21> i havent actually ever been asked for money by a stranger in melbourne proper
09:30:55 <kmc> maybe they just aren't very good
09:31:07 <kmc> or maybe i'm secretly racist :(
09:31:17 <shachaf> Clearly the optimal instrument should have one string.
09:31:28 <kmc> shachaf: yeah, i think it was that
09:31:45 <shachaf> kmc: I saw one played once and it seemed OK.
09:31:52 <shachaf> Then again, it was played by a violinist.
09:31:55 <itidus21> Phantom_Hoover: but yes... guys on street corners selling magazines that people only buy to support the people selling em
09:32:19 <itidus21> and buskers.. and otherwise people sitting and eating.. or bustling down the street
09:32:20 <kmc> the optimal instrument is the ondes Martenot
09:33:04 <itidus21> whenever you want to cross the road in melbourne you gotta watch out for trams
09:34:00 <itidus21> it's a prominent part of the experience
09:34:01 <kmc> you have to watch out for buses anyhere
09:34:11 <kmc> and trams are like buses except that they move more predictably
09:34:16 <kmc> though i guess they are quieter
09:34:29 <itidus21> kmc: but these trams they're really packed into a small area
09:34:37 <kmc> yeah trams can be longer too
09:35:02 <kmc> i rode a 6 car tram in budapest
09:35:08 <kmc> longest in the world
09:35:51 <kmc> going to go to lunch and then tallinn
09:41:21 <itidus21> i guess it's that the tram infrastructure makes them quite prominent.. with the rails, and the sheltered seating
09:43:00 <itidus21> from WP "The largest cable system in the world, in the city of Melbourne, Victoria, Australia, had at its peak 592 trams running on 74 kilometres of track."
09:43:50 -!- AnotherTest has joined.
10:23:57 -!- epicmonkey has quit (Ping timeout: 276 seconds).
10:39:10 -!- derdon has quit (Remote host closed the connection).
11:30:14 -!- ais523 has joined.
11:31:58 <ais523> you know what? Microsoft have some of the sanest terms of use and license agreements I've ever seen
11:31:58 <lambdabot> ais523: You have 1 new message. '/msg lambdabot @messages' to read it.
11:32:10 <coppro> except we aren't wuite ao allegic to government
11:32:19 <ais523> my guess is, because they get sued so often they have to make absolutely sure there's nothing in there that might bring the whole agreement down
11:32:23 <lambdabot> quintopia said 1d 12h 12m ago: I needed to know the "official" way of making colored text on mediawiki for Talk:BF_Joust_Strategies
11:33:05 <kmc> itidus21: cable-pulled cars?
11:33:14 <kmc> do they have those still?
11:35:21 <kmc> ah not since 26 October 1940
11:35:49 <kmc> SF is the only city to still run cable cars in normal traffic i think
11:35:50 -!- epicmonkey has joined.
11:54:10 -!- augur has quit (Remote host closed the connection).
11:55:44 -!- augur has joined.
11:56:42 -!- ais523 has quit.
11:56:51 -!- ais523 has joined.
12:00:01 -!- augur has quit (Ping timeout: 245 seconds).
12:00:27 <nortti> http://patriciopalladino.com/blog/2012/08/09/non-alphanumeric-javascript.html
12:01:21 -!- boily has joined.
12:03:23 -!- ais523 has quit (Read error: Connection reset by peer).
12:03:26 -!- ais523_ has joined.
12:03:42 -!- ais523_ has changed nick to ais523.
12:11:30 <kmc> not sure if dead pixel or unusual diacritic mark
12:29:06 -!- augur has joined.
12:39:52 -!- pikhq has quit (Quit: leaving).
12:55:33 <kmc> i wonder how secure my disk encryption passphrase should be
12:55:41 <kmc> the one i have now is quite long and annoying to type
12:57:34 -!- boily has quit (Quit: Poulet!).
12:59:25 <kmc> i'm mainly interested in keeping laptop thieves from being able to read my data
12:59:39 <kmc> and expect that a random laptop thief will not try extremely hard
12:59:47 -!- boily has joined.
13:00:00 <kmc> so i think it does not need to be very secure at all
13:02:31 <kmc> there's no reason it shouldn't just be my login password
14:02:42 -!- lexande has joined.
14:11:31 -!- Taneb has joined.
14:39:22 -!- TodPunk has quit (*.net *.split).
14:39:22 -!- atehwa has quit (*.net *.split).
14:39:22 -!- HackEgo has quit (*.net *.split).
14:39:22 -!- EgoBot has quit (*.net *.split).
14:39:22 -!- SimonRC has quit (*.net *.split).
14:39:22 -!- ssue has quit (*.net *.split).
14:39:22 -!- shachaf has quit (*.net *.split).
14:39:22 -!- comex has quit (*.net *.split).
14:44:58 -!- TodPunk has joined.
14:44:58 -!- atehwa has joined.
14:44:58 -!- HackEgo has joined.
14:44:58 -!- EgoBot has joined.
14:44:58 -!- SimonRC has joined.
14:44:58 -!- ssue has joined.
14:44:58 -!- shachaf has joined.
14:44:58 -!- comex has joined.
14:50:14 -!- derdon has joined.
15:07:49 -!- canaima172423 has joined.
15:08:47 -!- augur has quit (Read error: Connection reset by peer).
15:08:58 -!- augur has joined.
15:15:35 -!- canaima172423 has left.
15:19:03 -!- ais523 has quit (Ping timeout: 276 seconds).
15:20:10 -!- augur has quit (Ping timeout: 244 seconds).
15:20:12 <Gregor> esoteric = Español… oteric.
15:21:20 <itidus21> i suspect any word beginning with es is curious
15:22:15 <Gregor> That's just the establishment talking.
15:23:57 <itidus21> when i started googling for such words i encountered esne
15:24:01 <ion> Essentially yes.
15:24:07 <itidus21> which reminds me of a certain esolang
15:24:31 <ion> Googling? grep '^es' /usr/share/dict/american-english-huge
15:25:41 <itidus21> as for usr.. i suspect i have about 10 of them
15:27:06 <itidus21> the windows file hierarchy starts off good enough C:\Documents and Settings\
15:27:16 <itidus21> and at about that point its all fucked up
15:30:51 <itidus21> i freaked out because i got my graphics card on my birthday
15:31:08 <itidus21> and it was therefore a folder created on my birthday
15:39:24 <kmc> hello from Estonia
15:43:05 -!- augur has joined.
15:49:16 -!- elliott has joined.
15:49:29 <elliott> Who do I usually bother for git help again?
15:49:50 <kmc> no, bother #git and then talk smack about them behind their backs here
15:49:54 <kmc> that's the way
15:50:10 <elliott> That's like ~2.5x as much effort though.
15:50:45 <kmc> shachaf: I'm reading the thing you told me to read!
15:50:56 <kmc> lexande also told me to read it (separately)
16:15:21 * kmc can sometimes provide git help, as well
16:18:07 <elliott> Too bad, I solved it without the aid of anyone Finnish or called kmc!
16:21:38 <Taneb> elliott, if you ever have any problems with git again, just remember I am not the person you want
16:23:35 -!- zzo38 has joined.
16:29:06 <kmc> what was the question, out of curiosity
16:30:45 -!- epicmonkey has quit (Ping timeout: 248 seconds).
16:31:13 <kmc> www.quora.com/Palo-Alto-CA/Why-are-there-so-many-questions-about-Palo-Alto-CA
16:32:52 <elliott> kmc: the related questions are good too
16:33:12 <elliott> if i've merged something in a way that produces a merge commit and i oops'd and actually wanted it to be a rebasey merge so the merge commit was gone how can i do that
16:33:20 <elliott> the answer is start over using pull --rebase this time and perform the merge by hand again
16:33:33 <elliott> or use someone's shell script that trains this git rerere thing and i stopped reading at that point
16:35:16 <kmc> i would just reset to the point before the merge, and then cherrypick my work from after that
16:35:20 <kmc> which amounts to the same thing
16:36:38 -!- aloril has quit (Ping timeout: 244 seconds).
16:37:02 <kmc> i don't know what git rerere is
16:39:01 <lexande> elliott: is kmc an honourary Finn?
16:39:32 <Gregor> Everyone in this channel who isn't from Hexham is a Finn. Not honorary or otherwise, just a Finn.
16:40:31 <elliott> lexande: No. Someone is an honorary Finn, but I completely forget who.
16:40:56 <elliott> lexande: maybe it's you???
16:41:38 <Taneb> MC Hammer's not so good first draft of his song
16:42:17 <lexande> though I will be in Finland tomorrow
16:42:38 <Taneb> You are... American?
16:43:42 <kmc> "Stop!.... Continue!" is an Electric Six lyric
16:44:14 <kmc> from "Improper Dancing"
16:44:37 <elliott> kmc: By the way, why is make so bad?
16:44:48 <kmc> i don't have an answer to that
16:45:18 <lexande> are there better make replacements?
16:47:07 <Taneb> You know what annoys me?
16:47:27 <Taneb> On Facebook, it suggest articles. But to look at them, I have to let Yahoo post on my wall
16:48:18 <Taneb> Oh, that's a good idea
16:49:11 <kmc> what does tup have to do with git?
16:50:01 <elliott> It has nothing to do with git, but gittup has to do with both tup and git: http://gittup.org/gittup/
16:50:09 <kmc> TCP is not really working on this connection
16:50:47 <Taneb> lexande, the article was awful, why did I bother
16:50:48 <elliott> I'll just quote the best bit for you:
16:50:50 <elliott> "Yeah, I'm marf. Yeah, my computer is captainfalcon. Yeah, I just edited the spellcasting in nethack because I felt like it. Then I changed ls to print "Sup bro, way to list those files", because I like to be encouraged when I list things, and because I like it when people call me "bro"1. Then I re-compiled both and got a new initrd in like two seconds. Go ahead -- try to change the ls on your system to print out extra messages for no reason!2 Yo
16:50:51 <elliott> u can't do it!! Unless of course you're running gittup.org. But if you're running gittup.org, why aren't you playing nethack or needlessly recompiling things just for fun? In fact, how are you reading this webpage?? It doesn't even come with a web browser."
16:51:39 -!- aloril has joined.
16:52:17 <kmc> lockamee i'm zany and wacky
16:52:22 <kmc> bacon ninja bros
16:52:39 <kmc> i bet this person writes a lot of ruby
16:52:42 <lexande> elliott: does the http://www.jerkcity.com/jerkcity490.html strategy work with tup?
16:52:51 <elliott> gittup is cool and tup is cool and mike is cool and you are bitter
16:53:25 <elliott> lexande: there is someone with the nick bonghitz in a channel i am in
16:53:28 <lexande> kmc, the thing elliott pasted didn't seem all that formulaic
16:53:43 <kmc> elliott: i'm just taking the piss mate
16:53:52 <kmc> (been traveling with british people for 3 weeks, this is how i talk now)
16:54:03 <kmc> i think all four of those claims are likely true
16:54:14 <kmc> lexande: you are british when you are around actual british people
16:55:22 -!- madbr has joined.
16:56:44 <kmc> speaking of ninja
16:56:49 <kmc> i wonder how tup compares to ninja
16:57:01 <elliott> ninja is probably worse though
16:57:10 <elliott> because the building-the-build-system thing is kind of a pain in practice
16:57:21 <elliott> I have a patch in ninja because I wanted to use it to build Haskell code
16:57:23 <kmc> so why did the ninja developer(s?) decide tup wasn't adequate
16:57:27 <elliott> tup can't manage that right now
16:57:32 <elliott> because of the multiple output files
16:57:48 <kmc> multiple output files are not rare
16:57:55 <elliott> well, it's not so much multiple output files as the kind of
16:57:57 <kmc> latex and lex/yacc both come to mind
16:57:59 <elliott> pseudo-separate compilation GHC has
16:58:08 <elliott> where it's sort of separate compilation but not really
16:58:15 <kmc> what's special about it?
16:58:17 <elliott> kmc: anyway ninja was just intended to be incredibly fast and "trivial" I think
16:58:39 <kmc> ninja wants you to specify the dependency graph in a very explicit form
16:58:40 <elliott> despite this it still supports variables for no apparent reason
16:59:16 <elliott> let me try and find the discussion on the tup list
16:59:20 <nortti> gittup seems pretty awesome
17:00:17 <kmc> why do you say ghc's separate compilation is 'pseudo'
17:00:21 <elliott> kmc: here: https://groups.google.com/group/tup-users/browse_thread/thread/62eacaa9a22d9965/38468408c8b1aafb
17:00:28 <elliott> that will probably explain it better than I could over IRC
17:00:33 <elliott> but GHC does not have fully separate compilation
17:00:36 <elliott> because there are dependencies on interface files
17:01:59 <elliott> frankly the only reason I am not using tup for my current use case is because its output is quite ugly right now :p
17:02:12 <elliott> (it overdoses a bit on the ANSI colour)
17:02:35 <kmc> pipe to cat, problem solved
17:02:49 <elliott> i actually like its progress bars and some of the colour, it just overdoes it
17:02:50 <kmc> (surely it doesn't print color codes when stdout isnotatty, right?)
17:02:59 <elliott> I don't even know what it does when you pipe it to cat
17:03:13 -!- ogrom has joined.
17:03:58 <kmc> is there a particular reason why you consider "don't need to have compiled dependencies first" to be fundamental to the idea of separate compilation?
17:04:27 <elliott> that's more or less my definition of fully separate compilation :P
17:04:39 <kmc> yeah but is there some particular reason to take that definition
17:04:57 <kmc> it seems like a severe restriction
17:04:59 <elliott> well, something like C is distinctly more "separate" than Haskell in terms of how you can compile it
17:05:07 <elliott> it seems reasonable to have a term to distinguish the two
17:05:15 <kmc> and i'd rather have GHC's approach compared to C or ML where you're expected to duplicate information in a header file
17:05:27 <elliott> not saying fully separate compilation is the right trade-off
17:05:37 <elliott> but, relevantly, it makes building quite a bit simpler
17:05:47 <kmc> also you could write a haskell compiler which just parses imported sources and doesn't require the .hi
17:05:53 <kmc> but it wouldn't ddo things like cross-module inlining
17:05:56 <elliott> but it would be slow and bad
17:06:35 -!- madbr has left.
17:07:25 <kmc> if only ghc --make supported parallel builds
17:08:01 <kmc> Haskell and GHC are equal parts "advanced technology from the future" and "embarassingly behind the times"
17:08:04 <kmc> it's pretty amusing
17:08:28 <kmc> or you know, adjust the word "equal" to suit your attitude, i don't really care
17:08:29 <elliott> kmc: the idea was to make a tool which wrapped around cabal
17:08:29 <Phantom_Hoover> It's like how the Martians in War of the Worlds hadn't invented the wheel.
17:08:37 <elliott> by telling it "ghc" was actually my-ninja-using-ghc
17:08:42 <elliott> which would basically get the list of modules it gives
17:08:50 <elliott> make a ninja build file for them that uses non --make ghc
17:08:56 <kmc> Phantom_Hoover: evolution hasn't invented the wheel, either!
17:08:59 <kmc> there are no animals with wheels
17:09:01 <elliott> but then I lost interest and gave up : )
17:09:15 <kmc> what about them
17:10:00 <kmc> they roll along their axis?
17:10:13 <kmc> but they can't spin indefinitely, right?
17:10:14 <Taneb> That's more like a propeller, aren't they?
17:10:25 <kmc> like as many degrees as you like in one direction
17:10:41 <Taneb> They're used for a different purpose
17:11:22 <Taneb> But answer this...
17:12:03 <Taneb> ...nah, I can't think of a question
17:12:15 <kmc> propellers are wheels with bits put on
17:12:19 <kmc> OR DID I JUST BLOW YOUR MIND
17:12:30 <kmc> Phantom_Hoover: that's cool
17:12:40 <elliott> I just remembered make -jN is totally broken
17:12:52 <Taneb> Broken in your favour?
17:13:29 <Phantom_Hoover> kmc, it's also essentially a modified intercellular syringe.
17:13:33 <elliott> monqy: maybe we should switch to aimeka
17:14:24 <monqy> aside from a build system i mean
17:15:09 <elliott> monqy: the build system ais523 wrote for nethack 4
17:15:13 <Phantom_Hoover> just be careful that the ai makes don't rise up and start trying to unmake their human masters
17:15:14 <elliott> the idea is that you just run it and it figures out everything itself
17:15:19 <elliott> which means it works on nethack 4
17:15:22 <elliott> and doesn't seem to work on anything else
17:15:49 <elliott> also it's all one gigantic perl script
17:15:57 <monqy> 10:15:39 <monqy> beautiful
17:15:57 <monqy> 10:15:43 <monqy> we should use it
17:16:05 <elliott> i think i tried to build 41qys-crawl with aimake once but it broke
17:17:49 <Taneb> Does the ai stand for ais523?
17:18:39 <elliott> kmc: so if TCP isn't working are you using mosh
17:20:13 <elliott> note to self: @printf '\e[5;37;40;1m $(CXX) \e[0;37;1m $<\e[0m\n'
17:21:57 <zzo38> Do you know what way to convert to/from NTSC composite signals? Should ImageMagick have a command to do such things?
17:24:12 <elliott> to the second one, I mean!
17:25:10 <elliott> kmc: do you know any way to use make -jN and still get error messages that aren't lines and lines away from the compilation command
17:27:47 -!- Taneb has quit (Ping timeout: 244 seconds).
17:35:54 -!- epicmonkey has joined.
17:48:43 <olsner> one does not simply hi into #esoteric
17:49:33 <olsner> any day now, Gregor will implement the new rule that anyone who says hi will be kickbanned
17:49:57 <olsner> otoh, he might still be too busy admiring his pants
17:50:02 <Gregor> Implementing the rule is easy, but I don't have ops.
17:50:07 <Gregor> And yes, I do have pants.
17:50:11 <Gregor> Glorious, glorious pants.
17:50:38 * Gregor 's pants say, “Hello!”
17:51:19 <quintopia> i mean, i knew your pants were loud, just not articulate
17:51:59 <Gregor> They're pretty amazing pants.
17:52:15 <quintopia> what's that? couldnt hear that over your pants
17:52:24 <Gregor> http://www.solidcolorpants.com/servlet/the-3669/Purple-Dress-Pants-Trousers/Detail <-- this shall be next. Isn't this the most amazing pants color EVER?
17:53:29 <Gregor> They may advertise it as eggplant, but that ain't no eggplant.
17:53:47 <quintopia> the domain name suggests they might not be the most amazing color
17:54:02 <quintopia> the most amazing colors involve color combinations
17:54:39 <Gregor> quintopia: Although I would pay almost any price for pink and light blue angle-striped pants, I have something of a commitment to solid colors in my wardrobe as a whole.
17:55:53 <Gregor> Although I'm sure there are eggplants available in that color, I would say it's a very poor choice of color name.
17:56:39 <quintopia> no one will be able to hear you over EGGPLANTS FOR LEGS
17:56:49 <Gregor> I find this completely acceptable.
17:59:22 <boily> the new eggpants 2012! they burn through your retinas in 5 seconds flat!
17:59:47 -!- ais523 has joined.
18:00:05 <quintopia> would you not say that that color is precisely aubergine boily?
18:01:08 <boily> quintopia: yeah, but I wouldn't be able to make bad puns as easily as with "eggplant".
18:01:29 <boily> besides, hi quintopia!
18:01:29 -!- Taneb has joined.
18:01:42 <boily> (I say hi! in defiance of any vague kickban threats!)
18:07:47 -!- zzo38 has quit (Remote host closed the connection).
18:08:22 -!- nortti_ has joined.
18:11:01 -!- nortti has quit (Ping timeout: 248 seconds).
18:11:59 -!- nortti has joined.
18:15:02 -!- nortti_ has quit (Quit: Found terminal).
18:19:13 <Gregor> boily: There's no threat since I don't have ops ;)
18:24:16 -!- nortti has quit (Ping timeout: 240 seconds).
18:36:20 -!- nortti has joined.
18:40:11 <Phantom_Hoover> What about that campaign to give you voice, I was fully behind that.
18:40:48 -!- ogrom has quit (Read error: Connection reset by peer).
18:41:06 -!- ogrom has joined.
18:44:13 -!- Taneb has quit (Ping timeout: 244 seconds).
18:48:08 -!- oonbotti has quit (Remote host closed the connection).
18:50:01 <Sgeo> Gregor hates ops in rooms where everyone is given ops.
18:51:52 <Gregor> I have an aversion to normalcy.
18:52:08 -!- oonbotti has joined.
18:52:50 <Gregor> I hae no aversion to being given +v, it just seems unlikely is all :)
18:59:25 -!- augur has quit (Remote host closed the connection).
19:01:30 <Gregor> Luckily, my bright yellow pants are like an antinormalcy booster shot.
19:02:31 -!- Taneb has joined.
19:03:38 <Phantom_Hoover> Taneb, will you join my campaign to give Gregor voice.
19:11:42 -!- oonbotti has quit (Quit: oonbotti).
19:12:15 -!- oonbotti has joined.
19:15:21 <Taneb> Phantom_Hoover, what campaign
19:15:27 <Taneb> Why is Gregor mute?
19:15:47 <Taneb> GREGOR DRINK SOME HONEY WHISKEY MIXTURE!
19:16:30 <Taneb> Gregor, you can still drink some honey whiskey mixture, if you want to
19:16:34 <Phantom_Hoover> The one that lets you talk if +m is on (+m is never on).
19:16:36 <Taneb> And it's legal where you are
19:16:54 <Taneb> Phantom_Hoover, sure, I'll support your campaign
19:17:09 <FreeFull> voice also allows you to talk when banned
19:17:30 <Taneb> Is this across all IRC or just this channel?
19:17:48 <Phantom_Hoover> OK, I change the motion to: Gregor gets voice, banned.
19:19:09 <Taneb> Wow, Gregor is mute!
19:19:22 <Taneb> GREGOR DRINK THAT HONEY WHISKEY MIXTURE STAT
19:24:17 -!- ogrom has quit (Read error: Connection reset by peer).
19:28:56 -!- asiekierka has quit (Quit: Connection reset by PO).
19:39:39 -!- AnotherTest has quit (Quit: Leaving.).
19:43:30 <kmc> shachaf: Three Men in a Boat (To Say Nothing of the Dog)
19:43:50 <kmc> elliott: redirect stdout, maybe?
19:44:05 <kmc> elliott: yes, i'm using Mosh, and yes I know it's advertising ;)
19:47:02 <shachaf> kmc: No context in here, but did you read this story-thing by Ted Chiang? http://www.clarku.edu/welcome/placement/pdf/reading.pdf
19:47:12 -!- zzo38 has joined.
20:16:19 <kmc> shachaf: have read now
20:16:25 <kmc> where did you find it?
20:17:13 <shachaf> In a book, http://www.amazon.com/Stories-Your-Life-Others-Chiang/dp/1931520720
20:18:17 <kmc> books are neat
20:19:43 <kmc> the grpuabybtl gung znxrf fbzrbar va n ivqrb hahfhnyyl crefhnfvir is a plausible and frightening idea
20:20:19 <elliott> imo kdjfoiw owijer nfoaijg joiw owejf
20:20:55 <olsner> oh, is it rot13 or somesuch popular crypto?
20:22:44 <kmc> rot13, standard for hiding spoilers
20:22:57 <kmc> don't know what elliott is saying
20:23:48 <elliott> i smashed my keyboard randomly
20:23:52 <elliott> imo you should use rot14 instead
20:24:12 -!- liljuska has quit (Quit: Leaving.).
20:25:05 <olsner> to hide the spoilers from people who've learned to read rot13 by heart, of course
20:29:11 <kmc> elliott: that's pretty good random smashy
20:29:17 <kmc> in that i can't tell immediately what layout you use
20:31:53 <Taneb> Qwerty UK, I believe, although he was going to switch to Dvorak iirc
20:32:17 <olsner> but with above-average row switching for "random" smashing
20:32:29 <shachaf> kjio are contiguous, for instance.
20:32:46 <Taneb> hfawiulevghubaweyvgjuh
20:32:50 <shachaf> Of course I don't know any layouts other than QWERTY.
20:33:35 <shachaf> Good to know that I have a /usr/games/rot13
20:33:42 * shachaf wonders why games go in /usr/games.
20:33:56 <Taneb> Because you don't put games in the bin.
20:34:25 <olsner> rot13 is just tr 'a-z' 'n-za-m'
20:34:27 <kmc> apparently you can see finland from the tallinn tv tower
20:35:08 <elliott> i have no plans to switch to dvorak
20:35:09 -!- hagb4rd has joined.
20:35:21 <olsner> kmc: oh, that's almost transam
20:35:23 <elliott> ewofji oeorkf oisrjefj oijeoijewf jjfrkoa owjef
20:35:31 <elliott> basically i just rest my two hands and dance them simultaneously
20:35:40 -!- Taneb has changed nick to atriq.
20:35:46 <atriq> Hey, not registered
20:36:05 <kmc> elliott: you should switch to colemak, then you can be smug at dvorak and qwerty users
20:36:11 <kmc> i.e. the vast vast majority of people
20:36:15 <elliott> kmc: Deewiant uses Colemak!
20:36:20 <elliott> it seems like a pretty good layout
20:36:22 <kmc> but are they smug
20:36:27 <atriq> elliott,, think about this one
20:36:31 <elliott> Deewiant doesn't need to be smug
20:36:46 <kmc> does colemak come installed on major OSes?
20:36:51 <kmc> i think Ubuntu has it these days
20:37:01 <kmc> every OS has US-Dvorak since longtime
20:37:02 * shachaf wonders if the vast majority of people use keyboards.
20:37:11 <kmc> shachaf: maybe if you include mobile phones
20:37:20 <atriq> shachaf, I'd imagine people who use computers tend to use keyboards
20:37:29 <atriq> And most people I know use computers
20:37:37 <kmc> most people you know != most people
20:37:47 <olsner> the vast majority of people live in asia and make carpets, they have no need for keyboards
20:37:54 <atriq> But we must start somewherwe
20:38:03 <kmc> that's not really true
20:38:06 <hagb4rd> but they still need fridges
20:38:14 <kmc> poor people in asia still want to read news, watch porn, talk to friends on facebook like anyone else
20:38:22 <hagb4rd> especially in times of global warming
20:38:24 <shachaf> There are many more people in the world than there are IPv4 addresses.
20:38:31 <kmc> they may do this at internet cafes or on mobile phones rather than on computers they own
20:38:35 <olsner> I'd argue that touch keyboards don't count as keyboards
20:38:53 <olsner> ironically, it's impossible to touch type on a touch keyboard
20:39:52 <kmc> i found Pizza Americana in Tallinn
20:39:59 -!- epicmonkey has quit (Ping timeout: 244 seconds).
20:40:44 <atriq> AVOID IT LIKE THE PLAGUE
20:41:01 <kmc> do you have specific knowledge or are you just going from the name? ;)
20:41:10 <kmc> it would amuse me to eat at an america-themed restaurant while in foreign
20:41:27 <zzo38> I am one of the guys that doesn't want to watch porn and talk to friends on facebook like anyone else.
20:41:40 <atriq> Basically, avoid eating your own country's cuisine while abroad
20:42:03 <hagb4rd> you don't want to watch porn?
20:42:04 <atriq> atriq isn't a bad nick
20:42:13 <kmc> well most foods invented in America are nominally part of someone else's cuisine
20:42:20 <kmc> like pizza and orange chicken
20:42:36 <zzo38> hagb4rd: That is correct.
20:42:54 <hagb4rd> but you still have balls, don't you?
20:42:54 <atriq> hagb4rd, neither do I, but for probably different reasons
20:42:54 <kmc> i don't think there's an idea of "traditional American food" like the idea of "traditional Latvian food"
20:42:59 <kmc> unless you mean native american, of course :)
20:43:13 <atriq> I'm a raging, flamboyant asexual
20:43:31 * kmc wasn't sure whether the balls comment was about zzo38 or about food
20:44:12 <kmc> it's said that the most authentically British dish is chicken tikka masala
20:44:20 <zzo38> O yes, I have golf balls, pokeballs, and maybe some others which I have not found recently and forgot about.
20:44:32 <Phantom_Hoover> olsner, polenta is not actually food, it's a construction material that's categorised as food due to EU law.
20:44:41 <kmc> we went to a british irish-american pub in croatia which serves dutch beer and thai curry
20:45:07 <kmc> zzo38: but who's got the biggest balls of them all?
20:45:36 <hagb4rd> i guess the one who is not watching porn in the internet
20:45:51 <kmc> there is a paradox of authenticity
20:46:04 <kmc> when traveling do you go for the restaurant serving "traditional authentic cuisine" to tourists
20:46:08 <zzo38> kmc: You, isn't it?
20:46:14 <kmc> or do you find the place where locals eat, which probably serves kebab and pizza
20:46:27 <zzo38> (I don't actually know, I just guess)
20:46:30 <kmc> i live in boston and i almost never eat seafood
20:46:37 <kmc> zzo38: my balls are normal size, as far as i know
20:46:58 <atriq> kmc, I live in Hexham and rarely eat leather or chipboard
20:47:13 <kmc> maybe you should try!
20:47:32 <atriq> Actually, Hexham has a greek restaurant
20:47:39 <shachaf> I live in California and rarely eat oranges.
20:47:50 <olsner> atriq: wow! it really does have everything
20:47:51 <kmc> and you rarely smoke marijuana, california's #1 cash crop
20:47:54 <shachaf> Are oranges the state food of CA?
20:48:05 <atriq> olsner, except a...
20:48:07 <kmc> burritos and pot brownies
20:48:15 <kmc> dude, i could make, like, a... pot burrito
20:48:26 <hagb4rd> in baker, they have excellent seafood
20:48:53 * Phantom_Hoover realises atriq is Taneb after briefly entertaining the possibility of a third Hexhammer in the channel.
20:49:10 <shachaf> The friend I was visiting a friend yesterday seems to enjoy California's #1 cash crop.
20:49:12 <atriq> Phantom_Hoover, "atriq" is rot13 for ngevd
20:49:31 <Phantom_Hoover> You can buy Scotland's national food in Edinburgh, although I've never heard of anyone actually eating it.
20:49:38 <kmc> deep fried mars bars?
20:49:49 <atriq> olsner, I have been for ages
20:49:55 <atriq> I believe Gregor coined Ngevd
20:50:07 <olsner> what happens about half the time anyone here changes nicks is that one person disappears forever and some new person appears
20:50:31 <atriq> Based on my initials N. G. v D., the G standing for George after my great grandfather George Elliott Moscrop
20:50:32 <kmc> seafood is expensive and not tasty enough to justify the expense, i feel
20:51:06 <shachaf> Your great grandfather is named after elliott?
20:51:10 -!- boily has quit (Quit: Poulet!).
20:51:17 <atriq> shachaf, someone in the elliott clan
20:51:32 <atriq> Actually he got his middle name from his grandmother
20:51:56 <shachaf> The nick "Ngevd" vaguely reminded me of a desert, which I suppose was not the intention.
20:52:41 <atriq> My great-great-great grandmother (in two ways! oh no!) was called Jane Elliott
20:53:20 <Phantom_Hoover> it is why elliott is a stunted little midget and ngevd has a stupid moustache
20:53:22 <atriq> Actually, when this incest happened my ancestors lived in Newcastle
20:54:15 <atriq> I think the best thing to do to avoid incest is to meet your cousins before adulthood
20:54:24 <kmc> shachaf: you should open a dispensary in the Mission and serve pot burritos
20:55:19 <shachaf> kmc: Surely that already exists in SF somewhere.
20:55:22 <kmc> atriq: are you talking about the aversion to having sex with people you've grown up with?
20:55:28 <kmc> i think that doesn't work unless you live together all the time
20:55:34 <atriq> Also, to not change your children's religion away from that which is every other close relative, shortly before dying
20:55:40 <kmc> this was a problem for the kibbutzim though
20:55:47 <kmc> the ones with collective child-raising
20:56:17 <atriq> kmc, meeting before adulthood helps
20:56:55 <atriq> I'm not sure what the term us
20:57:21 <shachaf> http://www.thegreencross.org/ -- this whole business is more streamlined and convenient than I expected.
20:57:37 <atriq> Phantom_Hoover, there are at least two cases of incest in my ancestry
20:57:42 <atriq> I'm more inbred than Caligula
20:57:57 <elliott> atriq: I think "aromantic" is the standard terminology??
20:58:01 <atriq> (not actually much of an achievement)
20:58:13 <shachaf> Not to be confused with "aromatic".
20:58:15 <atriq> Phantom_Hoover, Caligula wasn't actually inbred
20:58:25 <elliott> shachaf: Well, atriq *is* smelly.
20:58:48 <shachaf> elliott: Have you met atriq?
20:58:54 <Phantom_Hoover> But also it's quite possible to be more inbred than, say, the child of siblings.
20:59:19 <atriq> shachaf, I have not mated with elliott
20:59:28 <Phantom_Hoover> See the family tree of Charles II of Spain, or that of Cleopatra, although that's not so much a tree as a column.
20:59:29 <kmc> take it to #esoteric-orgy
21:00:01 <kmc> shachaf: nice
21:00:03 <elliott> kmc: apparently i am not invited to that channel
21:01:08 <shachaf> kmc: There should be a channel where you give free half-baked advice to anyone who asks.
21:01:44 <shachaf> I guess this channel works for that purpose, though.
21:02:31 <ion> phantom_hoover: Haven’t you noticed that both sides of an issue are always “pro”?
21:02:35 <kmc> i don't think incestual relations between consenting adults should be illegal
21:02:38 <kmc> but it's still fucking creepy
21:02:39 <ion> Pro-incest vs. pro-childlove
21:02:47 <ion> whoops, brainfart
21:03:20 <monqy> elliott made a joke
21:03:21 <kmc> i'm always not getting things
21:03:21 <shachaf> elliott: Oh, was that what kmc meant?
21:03:30 <kmc> shachaf: most cities have marijuana delivery
21:03:33 <kmc> just no nice webpage for it
21:03:35 <Phantom_Hoover> kmc, turning literally any reference to incest into an annoying Lannister joke: pro or pro?
21:03:41 <ion> Pro-childlove vs. pro-innocence
21:03:52 <kmc> http://penny-arcade.com/comic/2012/08/10
21:04:03 <kmc> pro-gun vs. anti-penis
21:04:49 <Phantom_Hoover> kmc, is it just me or has there been a very marked drop in PA's quality over the past... month?
21:04:52 <elliott> monqy: it would be bad to have to prefix every entry in makefile.obj with $(OBJDIR)/ right
21:05:02 <monqy> elliott: is it avoidable
21:05:18 <monqy> is it cleanly avoidable
21:05:23 <monqy> does avoiding it make it better
21:05:51 <kmc> Phantom_Hoover: i would say more than a month
21:05:57 <kmc> it's been kinda meh for a while now
21:06:03 <kmc> but still better than most webcomics for sure
21:06:06 <kmc> at least most i've seen
21:06:16 <kmc> the art is still top-notch
21:06:21 <Phantom_Hoover> Yeah, but it's like it went from "kind of meh" to "really, really meh" very quickly.
21:06:29 <kmc> can you pinpoint the change?
21:06:30 <monqy> what's a most webcomics? don't answer that.
21:07:07 <shachaf> What's the Best Web Comic?
21:07:07 <atriq> elliott why are there so many elliotts in haskell
21:07:10 <kmc> maybe if i were 4 years older i would rage against penny-arcade instead of xkcd
21:07:19 <kmc> as far as something everyone likes and quotes and references but isn't actually that good
21:07:30 <shachaf> kmc: I don't like or quote it.
21:07:30 <Phantom_Hoover> The jokes from then have all been kind of "here's a silly thing, we have made a silly thing and this is funny".
21:07:55 <elliott> the startling realisation when you realise that penny arcade wasn't really ever very funny
21:08:03 <elliott> *startling moment? who knows. who knows.
21:08:04 <kmc> well some of those are references to video games you probably haven't played
21:08:08 <shachaf> I don't particularly dislike it either, it just never seemed to be worth starting to read.
21:08:44 <kmc> nah i think it was very funny for a while
21:08:48 <kmc> still think so
21:09:08 <kmc> xkcd was also very good but not for very long
21:09:15 <monqy> how good is a very
21:09:27 <kmc> i think PA did a better job of evolving out of the initial joke format
21:09:27 <elliott> xkcd was pretty good for longer than people give it credit for
21:09:48 <shachaf> elliott: Man, those spider comics.
21:10:00 <Phantom_Hoover> elliott, so you'd put the decay after the 400-500 mark most would?
21:10:41 <elliott> Phantom_Hoover: maintaining my idealised self-image requires that i do not consider questions like that worth answering
21:10:44 <kmc> xkcd was a great comic for a little while
21:11:08 <kmc> then it was an ok comic which was still regarded as the pinnacle of nerd culture
21:11:24 <ion> Ooh http://madebyevan.com/webgl-water/
21:12:00 <kmc> if there's one thing i love, it's giving untrusted javascript direct access to the most complicated kernel subsystem and most complicated peripheral device in my computer
21:12:04 <atriq> I'm really starting to warm to Cyanide and Happiness
21:12:06 <Phantom_Hoover> ion, wow, it's like I have a real bucket of water sitting on my desk.
21:12:17 <kmc> atriq: it's funny because awful things happen to people
21:12:23 <elliott> cyanide and happiness is an awful awful comic
21:13:08 <kmc> yeah it's pretty f. bad
21:13:32 * Phantom_Hoover wonders if you could use scientomantic resonancimation to make a decent wave in that thing.
21:13:32 <shachaf> TWIST: that was 30 seconds ago
21:13:55 <Phantom_Hoover> I might have liked it when I was a worse person but unfortunately I was otherwise engaged in being a bad person.
21:14:00 <kmc> the only long-running comic i can say for sure has not suffered a quality drop is...
21:14:12 <monqy> thank you jerkcity
21:14:20 <monqy> #1 reliable webcomic lifelong award
21:14:25 <shachaf> Has Cyanide and Happiness suffered a quality drop?
21:14:29 <shachaf> How can you drop from the bottom?
21:15:02 <kmc> when there is trap set up for you in every corner of this town
21:15:09 <kmc> and so you learn the only way to go is... underground
21:15:21 <atriq> It's slightly less dead baby now?
21:15:45 <monqy> now the baby is only half dead
21:15:50 <elliott> i am convinced, sounds great
21:15:55 <shachaf> Mostly dead is slightly alive, I hear.
21:15:59 <kmc> basing a regular comic on offensive / "shock" humor is a really dubious decision
21:16:14 <kmc> because, while the best examples of such humor are fucking hilarious, it kinda raises the bar for the minimum quality joke you can get away with
21:16:25 <monqy> what if there was a comic but instead of being a comic it just crawled shock sites and pulled from them
21:16:28 <kmc> i'm sure the top 20 C&H strips of all time are fucking hilarious
21:16:28 <elliott> i am not really sure how jerkcity could drop in quality
21:16:30 <monqy> "one million hits"
21:16:36 <elliott> since it is basically based entirely around not having any
21:16:36 <kmc> elliott: yeah exactly
21:16:43 <kmc> that's how i can say for sure
21:17:06 <kmc> yeah that's unfortunate
21:18:50 -!- oerjan has joined.
21:18:56 <Phantom_Hoover> You could argue that Homestuck hasn't dropped in quality but it's not gag-a-day, it's ultra-subjective, and it has large-scale ups and downs which do alter its entertainment value.
21:28:06 <Phantom_Hoover> It's essentially Randall Munroe's Picto-Blog as xkcdsucks described ages ago.
21:28:24 -!- elliott has left ("Leaving").
21:29:11 <atriq> Hey, elliott's gone
21:29:22 <atriq> Now it is time to make the important decision
21:29:31 <atriq> How much do I like the name "atriq"?
21:29:46 <kmc> motorola atriq 4g
21:29:55 <atriq> Phantom_Hoover, I already have
21:30:07 -!- Phantom_Hoover has changed nick to Prince_Charles.
21:30:30 <atriq> I've still got ettioll registered
21:30:59 <atriq> And RocketKSquirrel
21:31:06 <Prince_Charles> I *definitely* recall having Prince_Charles registered.
21:31:09 -!- Prince_Charles has changed nick to Phantom_Hoover.
21:31:55 <olsner> oh, who was rocketsquirrel?
21:32:34 -!- Gregor has changed nick to RocketJSquirrel.
21:33:00 <Phantom_Hoover> I would also have Vonlebio if it wasn't my black ops nick.
21:34:32 <oerjan> <Phantom_Hoover> Aww, they expire with disuse. <-- they gave several weeks warning last they did it
21:42:47 <Phantom_Hoover> Fair point, it wouldn't really be fair if I locked Prince Charles out of a nick that I never even use.
21:44:12 <oerjan> he probably goes as hrhpc or something
21:45:28 <tswett> What's a polynomial of the form (x^n - 1)/(x - 1) called?
21:45:47 <oerjan> hi, polynomial of the form (x^n - 1)/(x - 1)
21:45:57 <oerjan> RocketJSquirrel: no, quite easy actually
21:46:31 <oerjan> i cannot recall any name for it, though
21:46:42 <shachaf> tswett: When n is 1, it's called "1".
21:46:45 <Phantom_Hoover> It's called "fuck you STEP organisers for expecting me to factorise every polynomial of degree 3 or less in my head".
21:46:54 <oerjan> geometric sum perhaps?
21:47:53 -!- ais523 has quit.
21:48:53 <oerjan> geometric series fits sort of
21:49:09 <oerjan> except some may consider it implicitly infinite
21:49:27 <Phantom_Hoover> A polynomial which looks easier to factorise than it i-- wait it's totally one of the easy to factorise ones.
21:49:54 <tswett> Here we go. It's called an "all one polynomial", or AOP.
21:49:58 <shachaf> Phantom_Hoover tries to factor every polynomial he sees.
21:50:10 <tswett> Is x^n - 1 actually easy to factor, in general?
21:50:29 <tswett> I thought they were called cyclotomic polynomials. But they're not.
21:51:33 <tswett> If n is composite, it's easy to factor further. x^3 + x^2 + x + 1 = (x^2 + 1) (x + 1).
21:51:54 -!- atriq has quit (Ping timeout: 246 seconds).
21:52:32 <tswett> Hm. I'm pretty sure x^n - 1 factors into cyclotomic polynomials, and no further.
21:53:14 <Sgeo> Does BZFlag count as an FPS?
21:54:02 <oerjan> tswett: i'm not quite sure if the cyclotomic polynomials are always irreducible?
21:54:04 <tswett> If it's got a first-person perspective, and you can shoot from it, then I think so.
21:54:26 <oerjan> oh gauss has proved that
21:54:50 <Sgeo> tswett, you can't look up or down. Also, you're a tank. But yes.
21:55:36 <kmc> you can't look up or down in doom
21:56:11 <Phantom_Hoover> (There are a couple of planes with working guns for FG.)
21:59:26 <oerjan> tswett: wp's cyclotomic polynomial article has a pretty simple way of calculating the cyclotomic polynomials which also factors x^n - 1 in the process
22:00:00 <oerjan> uses that x^n - 1 is the product of the d'th cyclotomic polynomial for d a factor of n
22:00:23 <kmc> shachaf: a bunch of the original ksplice people are quitting ksploracle
22:00:29 <kmc> this is... not surprising
22:02:53 <oerjan> 18:29:45: <Phantom_Hoover> Gregor, didn't you have ops for a while.
22:02:53 <oerjan> 18:36:03: <Gregor> Nope.
22:03:06 <oerjan> actually he did, for about a day or so
22:03:41 <oerjan> it was only +o ops though, not chanserv access
22:06:38 -!- sebbu has quit (Quit: ZNC - http://znc.sourceforge.net).
22:08:04 <oerjan> <kmc> apparently you can see finland from the tallinn tv tower <-- now i'm imagining an estonian sarah palin
22:10:06 <RocketJSquirrel> I love that people barely remember the fact that Palin never said that :)
22:10:29 <oerjan> poor michael, demoted to second craziest palin
22:11:45 <oerjan> RocketJSquirrel: larger than life people don't get to determine their own life facts
22:11:57 -!- nooga has quit (Ping timeout: 244 seconds).
22:15:01 -!- sebbu has joined.
22:16:22 -!- nooga has joined.
22:19:21 <oerjan> <atriq> Phantom_Hoover, "atriq" is rot13 for ngevd <-- darn i got fooled by the at -> ta switch and thought it would be rot 19 or something for taneb. wait that doesn't work for t -> a either...
22:20:18 <oerjan> i assume even RocketJSquirrel doesn't want to go for the green lime ones.
22:20:47 <RocketJSquirrel> oerjan: They were on my list 'til I saw the “eggplant” ones.
22:22:49 <shachaf> kmc: Did some vesting thing just end or something?
22:24:10 <shachaf> "Oracle Buys Ksplice", "Jul 21 2011"
22:24:35 <kmc> yeah some of them got some money after 1 year
22:25:03 <shachaf> So now there are a bunch of free-floating kspliceurs.
22:25:08 <kmc> others are leaving because their friends are, i imagine
22:25:13 <kmc> there is a coordination problem
22:34:33 -!- nooga has quit (Ping timeout: 276 seconds).
22:40:43 -!- david_werecat has joined.
22:45:57 -!- zzo38 has quit (Remote host closed the connection).
22:48:12 -!- augur has joined.
22:51:26 -!- david_werecat has quit (Ping timeout: 252 seconds).
22:53:16 <FireFly> http://www.kickstarter.com/projects/project-giana/project-giana
22:54:11 * oerjan swats FireFly -----###
22:55:10 -!- augur has quit (Remote host closed the connection).
22:55:45 -!- augur has joined.
22:55:51 <tswett> "So in short, what do moths eat? They eat anything that can dissolve in water."
22:55:58 <oerjan> Phantom_Hoover: i note how the sidebar has several other similar sites
22:56:43 <tswett> "Some images taken from Wikipedia under GNU"
22:57:42 <tswett> That's how free licenses work, right? "You are free to distribute this work, as long as you mention a couple of free-works projects with prepositions before them."
22:59:48 <shachaf> "under GNU" surely refers to "Among the hunting trophies on the wall above my bed / Stuffed and mounted, was a face I thought I knew"
23:00:08 -!- augur has quit (Ping timeout: 245 seconds).
23:00:56 <shachaf> I think F+S are pretty famous.
23:07:29 -!- david_werecat has joined.
23:08:15 <soundnfury> nor am I in the least like that dreadful hartebeest
23:14:25 -!- david_werecat has quit (Read error: Connection reset by peer).
23:14:46 -!- david_werecat has joined.
23:17:07 <oerjan> what girl genius is still not updated
23:21:06 -!- augur has joined.
23:22:14 <oerjan> i recall it used to be up shortly after 05:00 UTC, but lately it seems to update at an irregular time
23:25:38 -!- augur has quit (Remote host closed the connection).
23:29:23 -!- augur_ has joined.
23:37:03 -!- stanley has left ("Leaving").
23:42:22 -!- hagb4rd has quit (Ping timeout: 265 seconds).
23:48:37 <oerjan> did google's favicon just change?
23:48:44 -!- zzo38 has joined.
23:50:00 -!- augur_ has quit (Read error: No route to host).
23:50:37 -!- augur has joined.