00:08:45 -!- Gregor has quit (Ping timeout: 252 seconds).
00:30:24 <Ilari> Heh... From wisdom of crowds stuff: Task 1k random people to develop new nutrional guidelines and then do study of what those do to health. Wheither one would or would not want to see that depends on relative priorities of caring for others vs. how much one wants to see train wrecks... :-)
00:33:43 <Ilari> Because those recomendations would probably be total garbage (difficult to say better or worse than current official ones)...
00:39:11 <Sgeo_> What's wrong with current guidelines?
00:40:46 <Ilari> Some would say 'too much carbohydrates', but I won't...
00:43:10 <Ilari> Worst part: They are almost impossible to follow without snacking on garbage...
00:46:15 <Ilari> Also that the recomendations do not seem to be based on reality...
00:47:25 <Ilari> Did you know that US nutrional recomendations are published by USDA (agriculture) and not by HHS (Health)?
00:49:05 <Ilari> (well, latest ones are in co-operation with HHS, but one can still tell where the priorities lie...)
00:49:45 <pikhq> The old recommendations were, what, 11 servings of grains each day?
00:50:19 <Ilari> The really difficult questions: What causes metabolic syndrome? There are some suspect causes, some known not to be cause. But no unified model.
00:50:31 <pikhq> Yup, 6-11 servings.
00:51:19 <pikhq> I cannot fathom eating that much grain.
00:52:22 <pikhq> "Eat 5 times as much grain as meat, poultry, fish, beans, eggs, or nuts!"
00:53:08 <pikhq> Hmm. What was it, 1 slice of bread = serving? "Eat a loaf of bread each day!"
00:53:36 <Ilari> Fun "paradoxes" ("paradox" in food & health really means "we have the wrong model"): Whole milk is less fattening than skim milk. Epidemilogical studies pick association, animal studies pick causation.
00:54:49 <Ilari> Yeah, you need to eat lots of grains, since grains are nutrient-poor (that's for white grains, whole grains can be even worse).
00:55:41 <pikhq> It's also comical that they stuck pretty much everything with protein together...
00:56:05 <pikhq> Because apparently tofu and a steak are nutritionally equivalent.
00:57:43 <Ilari> Fun... Its impossible to satisfy both nutrional guidelines and what I think are healty macronutrient ratios (based on what has been observed in hunter-gatherers). Actually, only fundamental conflict is in saturated fat sector...
00:58:24 <pikhq> Not to say that the average US diet is any saner, but the government recommendations are fucking nuts.
00:58:44 <Ilari> (and I didn't refer to its taste).
00:58:49 <pikhq> "Now, eat a loaf of bread, but heaven forbid that you eat more than two pieces of fruit a day."
00:59:19 <pikhq> Tofu is not bad when prepared well. (which it almost never is on this continent)
01:00:27 <pikhq> (flavor-wise, that is. Nutritionally, you're looking at protein and not much else.)
01:00:31 <Ilari> Errr... Weren't fruits the second overhyped food group? Those tend to be quite high in sugar (of the apparently harmful kind)...
01:00:46 <pikhq> They said "2 a day".
01:01:16 <pikhq> Same as meats. And dairy products.
01:01:27 <pikhq> Vegetables were 3-5 a day.
01:01:55 <pikhq> They also said "use sparingly" on... All fats. *All* fats.
01:02:16 <Ilari> Oh, at least new recomendations have all sugar <10E%. Here sugar is <25E% (insane).
01:02:32 <Ilari> What's serving in dairy products?
01:03:16 <pikhq> 1 cup of milk. (the measurement "cup")
01:05:12 <Ilari> Ah, I would exceed that by quite a margin... :-)
01:05:44 -!- BeholdMyGlory has quit (Remote host closed the connection).
01:07:10 <Ilari> Meats are good proein sources...
01:11:43 <Ilari> Where does that fear of cholesterol come from? Not even Keys (who IIRC started the whole 'saturated fat is bad'/'high cholesterol causes heart disease' crock) thought that dietary colesterol is important (unless you happen to be rabbit or hamster)... :-/
01:18:56 <Ilari> Heh... One (legendary) doctor had 4 patients that wanted to gain weight. So he had them drink 100g of olive oil a day... Didn't work...
01:21:44 <Ilari> One with wrong model of things would call this "paradox"... :-)
01:28:33 -!- coppro has quit (Ping timeout: 240 seconds).
01:37:08 -!- coppro has joined.
01:39:44 -!- Mathnerd314 has joined.
01:47:10 <Sgeo_> Ilari, what makes you qualified to talk about nutrition, and others unqualified? Just asking..
01:49:30 <pikhq> Knuth has proposed a successor to TeX.
01:49:34 <pikhq> The syntax is XML.
01:49:52 <Ilari> Not that I'm qualified... There are others way more qualified than me. But there are also others (includin "experts") that are way less qualified.
01:50:11 <pikhq> XML is NOT a human-usable syntax.
01:50:15 <coppro> pikhq: Obviously it should be full SML
01:50:27 <pikhq> TeX is more so than XML.
01:50:38 <pikhq> Which is not to say that it's very usable, but... Ugh.
01:55:56 <coppro> okay, it's pretty obvious by now that was a joke
02:08:10 -!- Gregor-L has joined.
02:09:09 -!- oerjan has joined.
02:09:16 <coppro> is there some website where you can bet on how far Knuth will get on TAOCP before he croaks?
02:09:38 <coppro> it's quick on the way to being the greatest unfinished work of computer science
02:10:16 <pikhq> And still the greatest work.
02:11:07 <coppro> (I mean, yes, it is currently the greatest unfinished work, but you know what I mean)
02:13:18 <oerjan> 11:43:05 <oklopol> i wish oerjan was here
02:13:18 <oerjan> 11:43:31 <oklopol> are you talking about receiving head
02:23:41 -!- Gregor-L has changed nick to Gregor.
02:27:12 <oerjan> wyde is better than tyght
02:51:41 -!- cal153 has quit (Ping timeout: 240 seconds).
03:08:30 <coppro> Does sideways add serve any direct arithmetic use or something?
03:10:00 <coppro> an instruction on some architectures that takes two operands; the result has each bit set iff the first operand had it set and the second did not
03:11:05 -!- cal153 has joined.
03:11:09 <coppro> it's count the number of bits that meet those conditions
03:11:28 <oerjan> pikhq: er, i feel a draft over my head
03:12:05 <coppro> in MMIX (the spec for which I'm reading), it could serve to count bits set because there's no direct operation for that (just put it against an all-0 second operatnd)
03:13:07 <pikhq> Uh, isn't that with 0x0 on the second operand the identity function?
03:13:39 <coppro> like I said; I misspoke earlier
03:13:45 <coppro> it's a bit-counting instruction
03:13:56 <coppro> AND NOT against 0 is the identity operation, yes
03:14:03 <coppro> it's "count the bits in X AND NOT Y"
03:15:21 * oerjan doesn't know why one would want that particular combination often enough not to just construct it from simpler parts
03:16:08 <pikhq> oerjan: Blame Knuth.
03:16:28 <coppro> it's not just Knuth; I researched it and it's existed physically before
03:17:07 <oerjan> well, must be useful then :D
03:17:14 <pikhq> Well, he *did* try to make MMIX a CPU ISA that one could actually use.
03:17:35 <coppro> I imagine he has some use in mind
03:18:19 <oerjan> http://cryptome.org/jya/sadd.htm
03:20:28 <oerjan> hm that page so far doesn't really imply that it did any AND NOT stuff before counting
03:25:31 <coppro> whoa, the matrix operations are cool
03:27:02 <oerjan> that string search use of sideways add looks interesting
03:27:12 <oerjan> (but uses AND, not AND NOT)
03:29:03 <Ilari> Population count unit... Geez...
03:29:51 <Ilari> Very exotic instruction and optimized to ridiculous degree...
03:32:00 <Ilari> Haha: "Seriously, the milk section includes transparent cheese. Is that a new Kraft product?".
03:33:40 <Ilari> (referring to proposed 2010 dietary guidelines).
03:39:02 <coppro> I like how Knuth's fake architecture is designed with several instructions that no one would really want on a computer used only for educational purposes
03:41:14 <pikhq> It's not intended only for educational purposes.
03:41:43 <pikhq> Well, it is, but it's intended to be real-world usable so you actually learn something other than a bizarre educational subset of things.
03:42:09 <coppro> but one instruction is described as "reserved for operating systems only"
03:42:38 <coppro> "details are in MMIXware", which is not part of TAOCP
03:43:31 <Sgeo_> Fake architecture?
03:45:52 <coppro> Sgeo_: in that it has no hardware implementations
03:46:30 * Sgeo_ wants to see an arcitecture that can only be simulated
03:46:36 <pikhq> Though it certainly could.
03:46:48 <pikhq> It's actually a quite reasonable RISC architecture.
03:47:06 <pikhq> From Knuth and the guys responsible MIPS and Alpha.
03:47:43 <coppro> yeah, he basically designed it so that you could put Linux on it
03:47:45 <coppro> nice: http://www-cs-faculty.stanford.edu/~knuth/diamondsigns/CP6.html
03:48:00 <coppro> (note: I don't mean that he actually made that a design goal, merely that it is sufficiently comprehensive to allow that)
03:48:45 <pikhq> In other words: he didn't want it to be a toy.
03:49:14 <coppro> it even has compare-and-swap
03:50:28 <Ilari> Well, Compare-and-swap is absolute basic operation of multi-CPU operation.
03:51:04 <Ilari> IIRC, from that one can construct all manner of atomic operations and all basic multithread synchroninzation primitives.
03:51:42 <Ilari> What kind of odd instructions?
03:52:40 <Ilari> coppro: What instructions no one would really want on real computer?
03:52:52 <oerjan> KILL-THEN-ASK (for handling zombie processes)
03:53:07 <coppro> Ilari: uh, the sideways add is a bit weird
03:53:43 <Ilari> coppro: Sideways add? You mean population count?
03:54:05 <coppro> Ilari: it is defined of "the number of bits that are 1 in $Y and 0 in $Z"
03:54:39 <Ilari> Well, that's somewhat odd.
03:55:06 <Ilari> Well, one can use that as population count. Especially if MMIX has RISC-style zero register.
03:56:00 <coppro> Every operation has a form taking a constant as the third operand $Z, so yes, you can use it against 0 in one instruction
03:56:55 <coppro> also, internal interrupts involve jumps to low addresses
03:57:07 <coppro> (internal being within a program; not an OS-level interrupt)
03:57:41 <Ilari> What kind of interrupts there are within programs?
03:58:33 <coppro> conditions like integer overflow, divide by zero, or the like. Whether they are interrupts or just set flags is controllable, but if they are set as interrupts they all jump into the first few bytes of memory
03:58:55 <coppro> also, the TRIP instruction, which causes a manual interrupt, seems weird
03:59:10 <coppro> for starters, it involves a jump to address 0
03:59:36 <coppro> (TRAP, which is an external interrupt, is by contrast quite useful and normal)
03:59:42 <Ilari> I would want first page not to be mapped...
03:59:54 <pikhq> Yeah well screw you.
04:00:14 <coppro> Ilari: it does not have paging AFAICT
04:00:31 <pikhq> coppro: It does, but you're running in userspace so YOU CAN'T TELL.
04:00:41 <coppro> pikhq: Well, a kernel could implement it on top of the OS
04:00:48 <coppro> on top of the architecture
04:00:52 <coppro> but there's no architecture support
04:01:30 <pikhq> Someone made an MMIX simulator that ran unhosted code for the sole purpose of running Linux on it.
04:01:35 <pikhq> That sucker's got virtual memory.
04:01:59 <coppro> hmm... actually, it may
04:02:05 <coppro> there are some features left undescribed
04:02:08 <coppro> I don't think so though
04:02:13 <coppro> there isn't enough room in the instruction table
04:03:31 <coppro> stop hiding features in sideways comments, Knuth!
04:03:31 <Ilari> What all is segment descriptor in X86 is pretty crazy. There are ordinary code and data segments, but there's also TSSes, task gates, interrupt gates, exception gates, LDTs, etc...
04:04:07 <coppro> there's a reason CISCs are going out of styles
04:04:23 <Ilari> TSSes: Hardware-assisted task swapping...
04:04:34 <pikhq> Arguably, CISC is dead outside of 8-bit and 16-bit CPUs.
04:05:06 <pikhq> After all, modern x86 implementations are not CISC. They are RISC chips running a very fast x86 emulator.
04:05:07 <coppro> the no-op for MMIX is entitled "SWYM", or "sympathize with your machinery"
04:05:20 <Ilari> Yes, LDT is segment descriptor you stick into GDT. And some segment selectors reference entries in LDT.
04:06:06 <coppro> pikhq: yeah, you know things suck when...
04:07:02 <Ilari> Oh, and of course call gates.
04:08:45 <coppro> I'd have to look into MMIXware to see if TRIP has a real use
04:17:32 <Ilari> Oh, and x86 has expand-down segments, where segment limit is minimum address and not maximum address. As for why Linux/x86 doesn't set kernel CS to be expand-down: Expand-down code segments are not supported!
04:20:16 <pikhq> Oh, x86's segmentation.
04:20:33 <pikhq> If you were crazy enough, you could use it *and* paging for virtual memory.
04:24:25 <Ilari> System segment types: TSS (16 and 32 bit, available and busy variants of course), Call gates, Interrupt gates and trap gates (all with 16- and 32-bit variants).
05:46:18 -!- augur has quit (Remote host closed the connection).
05:46:49 -!- augur has joined.
06:01:24 <Gregor> "If you value your independence and creativity, you should be aware that Apple doesn't. Take your computing elsewhere." -- FSF
06:02:03 <coppro> maybe if we get the FSF and Apple into a major battle, they'll destroy each other and the world will be a better place
06:02:54 <lifthrasiir> coppro, are you sure that equivalent organizations (or company or so) won't arise after them? ;)
06:03:15 <coppro> lifthrasiir: there will be a Calm
06:04:23 -!- SgeoN1 has joined.
06:07:40 -!- clog has quit (Ping timeout: 245 seconds).
06:07:40 -!- clog has quit (ended).
06:07:43 -!- clog has joined.
06:07:43 -!- clog has joined.
06:19:13 <pikhq> coppro: So, you are proposing the plot of FFX as a solution.
06:20:49 <pikhq> All fun and games until some punk decides to kill off everything to spare people pain.
06:21:37 <coppro> uh, I think you're confusing games
06:22:30 <pikhq> Ending the spiral of death?
06:23:05 <coppro> oh sure. I thought you were talking about Yu Yevon
06:23:25 <pikhq> Nah, Yu Yevon's a different kind of crazy.
06:23:51 <pikhq> All fun and games until some punk decides to turn people in statues to run a massive summoning.
06:24:26 <coppro> and kill off everything
06:25:02 -!- oerjan has quit (Quit: leaving).
07:14:29 -!- SgeoN1 has quit (Quit: Bye).
07:27:49 -!- BeholdMyGlory has joined.
07:29:10 -!- coppro has quit (Read error: Connection reset by peer).
07:29:36 -!- coppro has joined.
07:45:16 -!- Phantom_Hoover has joined.
07:56:01 <fizzie> The twitterverse seems to say that it was a jokey thing "announcing" TeX's successor with XML syntax.
07:57:10 <fizzie> And then something about ringing a bell when you say the name.
07:57:19 -!- Phantom_Hoover has quit (Quit: Leaving).
07:57:32 -!- Phantom_Hoover has joined.
07:57:56 <fizzie> It ended up in Wikipedia for ~2 hours, too: http://en.wikipedia.org/w/index.php?title=Itex&oldid=371122107
07:58:00 <fizzie> Then again, everything does.
07:58:33 <coppro> it features 3-d printing and VP8
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:02:07 -!- Phantom_Hoover has quit (Client Quit).
08:02:07 -!- Phantom_Hoover has joined.
08:02:39 -!- pikhq has quit (*.net *.split).
08:02:39 -!- lifthrasiir has quit (*.net *.split).
08:02:58 <fizzie> Gah, if I tell mutt to search for a string in message bodies on this 1776-email IMAP inbox, it will fetch all message bodies. Wasn't there some sort of search functionality in IMAP? (Come to think of it, it's probably just that it wouldn't support regexps that way.)
08:04:59 -!- pikhq has joined.
08:04:59 -!- lifthrasiir has joined.
08:05:43 <fizzie> (And I didn't even find what I was looking for.)
08:07:09 -!- olsner has quit (Ping timeout: 265 seconds).
08:07:11 <AnMaster> fizzie, so Knuth's announcement was a joke?
08:07:26 -!- Phantom_Hoover has quit (Quit: Leaving).
08:07:32 -!- olsner has joined.
08:07:42 -!- Phantom_Hoover has joined.
08:08:22 <fizzie> AnMaster: So it appears.
08:08:55 <augur> he didnt actually make an announcement, did he
08:09:43 <fizzie> He made a joke at the end of a conference; it's not so uncommon, though usually the jokes don't come with this much pre-excitement.
08:09:55 <augur> what was the joke?
08:10:24 <fizzie> That there'll be a TeX successor called iTeX with XML syntax. It's just a dozen comments backwards on this channel, you know.
08:11:12 <augur> hes a bit weird tho
08:11:32 <augur> i mean, hes quite smart and knowledgable, but at the same time he seems to have a weird view of how numbers work mathematicall
08:12:20 <augur> tho he might be a strict formalist, in that regard
08:13:20 <augur> well, in one of his books, he made some comments regarding natural numbers and integers and so on and so forth and i wrote him saying that mathematically speaking, these two things were identical in some relevant fashion
08:13:38 <augur> er, well, natural numbers/integers and reals
08:13:46 <augur> basically involving the naturals/integers being a subset of the reals
08:13:51 * Phantom_Hoover wonders why one could possibly want an enum larger than an int
08:14:14 <augur> and he replied that you'd have to define an equivalency between them in order to make such and such blah blah blah
08:14:30 <fizzie> Phantom_Hoover: With 16-bit ints, you might only have room for puny 65k named constants.
08:14:41 <augur> yeah, i was saddened
08:14:57 <Ilari> Heh... There's Damn Vulernable Linux... Linux distro stuffed with outdated software, exploitable software and ill-configured software... The list of default services is probably impressive.
08:15:01 <augur> especially that it wasn't a simple enough error
08:15:30 <augur> like, if it were an error that could be made by simply forgetting to copy some symbol or other, or whatever, ok sure fine whatever
08:15:40 <augur> whoops, typo, so to speak
08:15:52 <augur> but this was kind of crucial to the point he was making
08:15:52 <Phantom_Hoover> fizzie, this was from the x86-64 ABI spec; it uses 32-bit ints.
08:16:32 <fizzie> When you want to name each and every memory location with a descriptive name?
08:17:25 <fizzie> Yes, so you need more than 32 bits for it.
08:18:11 <Phantom_Hoover> Wouldn't the compiler die of overwork trying to keep track of them?
08:19:03 <fizzie> Well, if you have a limited amount of cases but they're represented by bit-patterns wider than int, maybe?
08:19:36 <fizzie> A 64-bit register with few flags up high, and you want the values in an enum so that you can say "blah | bleh" to get a proper value.
08:19:40 <fizzie> Or something like that.
08:20:20 <fizzie> After all, it's allowed for the values of enum constants to be non-arbitrary integers.
08:20:27 <fizzie> fungot: You guess you what?
08:20:28 <fungot> fizzie: something like that, change the environment properties/ colors/ positions once a while back he snagged a bit of plot creativity, they could be
08:22:48 <fungot> Available: agora alice c64 ct darwin discworld europarl ff7 fisher ic irc* jargon lovecraft nethack pa speeches ss wp youtube
08:23:27 <fizzie> Where do you have wider-than-int enums, though?
08:24:47 <Phantom_Hoover> C++ and some implementations of C permit enums larger than an int. The underlying
08:24:47 <Phantom_Hoover> type is bumped to an unsigned int, long int or unsigned long int, in that order.
08:25:58 <coppro> Ilari: my first actions if I heard someone was running it: ssh root@box
08:27:32 <Ilari> coppro: You mean 'telnet box'? :->
08:27:56 <coppro> if that doesn't work, I'd try mysql
08:28:22 <Ilari> Actually, ssh might be better target, but there's no telling what they have used as in.telnetd (probably something really vulernable).
08:30:11 <Ilari> And not only vulernable, but misconfigured to create additional vulernabilities.
08:30:23 <Ilari> Oh, and the SSH probably has SSH1 enabled...
08:31:07 <coppro> yeah, but if root is passwordless or has a weak password, SSH wins (assuming it allows root logins at all, which is always a bad idea)
08:31:41 <Ilari> Hey, guess twice if it allows direct root logins? :-)
08:32:44 <Ilari> Well, judging from goals, it very probably does.
08:34:07 <coppro> I could see an instructor running that and giving bonus marks for every different way someone students came up with to root it.
08:35:19 -!- coppro has quit (Remote host closed the connection).
08:50:46 -!- MigoMipo has joined.
09:26:34 -!- MizardX has joined.
09:38:38 -!- BeholdMyGlory has quit (Ping timeout: 240 seconds).
09:43:18 -!- sebbu has quit (Ping timeout: 240 seconds).
09:43:29 -!- BeholdMyGlory has joined.
09:43:29 -!- BeholdMyGlory has quit (Changing host).
09:43:29 -!- BeholdMyGlory has joined.
09:45:10 -!- sebbu has joined.
10:14:38 -!- tombom has joined.
10:35:20 -!- GreaseMonkey has quit (Quit: I'm using NO SCRIPT WHATSOEVER - Download it at file:///dev/null).
10:46:29 -!- pikhq has quit (Read error: Connection reset by peer).
10:53:16 -!- pikhq has joined.
11:11:46 <Phantom_Hoover> AAGH the number 168 is stuck in my head and I don't know why.
11:23:55 -!- hiato has joined.
11:35:56 <AnMaster> hm I wonder if linux is smart enough when dealing with reading from software RAID1 that it uses the disk that happens to have the read head closest to what it needs to read?
11:41:51 <AnMaster> Phantom_Hoover, or did you mean which command line tool you use to set it up?
11:52:10 <fizzie> It sounds somewhat tricky to arrange it completely optimally; preferrably you'd start reading with the drive that has the least seek time to the target region, and then read in parallel from all drives as soon as they have reached the proper place.
11:53:10 <fizzie> I'm not even sure if you can get precise enough geometry information from disk controllers nowadays. And there's sector remapping and all that fluff.
11:55:59 -!- yiyus has quit (Ping timeout: 240 seconds).
12:11:05 <fizzie> From the "this howto is deprecated" Linux Software-RAID HOWTO, on the topic of RAID-1: "Read performance is good, especially if you have multiple readers or seek-intensive workloads. The RAID code employs a rather good read-balancing algorithm, that will simply let the disk whose heads are closest to the wanted disk position perform the read operation. Since seek operations are relatively expensive on modern disks (a seek time of 6 ms equals a read of 123 kB at
12:11:05 <fizzie> 20 MB/sec), picking the disk that will have the shortest seek time does actually give a noticeable performance improvement."
12:11:22 <fizzie> So at least at some point it has attempted to handle that cleverly.
12:11:37 -!- yiyus has joined.
12:15:27 <fizzie> The same phraseology (with updated numbers: 8 ms, 640 kB at 80 MB/sec) appears in raid.wiki.kernel.org, but on the other hand in the "Performance" page the benchmarks don't show very much read-speed differences between non-raid and RAID-1 access. (The benchmarks look somewhat haphazard, though.)
12:35:30 -!- pikhq has quit (Read error: Connection reset by peer).
12:40:26 -!- rodgort has quit (Quit: Coyote finally caught me).
12:40:34 -!- rodgort has joined.
12:43:28 -!- oerjan has joined.
12:45:48 <oerjan> but can you _truly_ be sure?
12:46:26 <Phantom_Hoover> No, but for all I know you could just be an emergent phenomenon of the intenet itself.
12:46:35 <oerjan> <fizzie> That there'll be a TeX successor called iTeX with XML syntax. It's just a dozen comments backwards on this channel, you know.
12:47:35 <oerjan> the very name of that + the preannouncements would seem to imply knuth was deliberately making a joke on apple hype
12:48:00 <oerjan> well except apple doesn't make preannouncements, do they
12:48:44 -!- hiato has quit (Quit: Lost terminal).
12:50:53 -!- hiato has joined.
12:52:09 <fizzie> They just do strictly controlled leaks with both correct and incorrect information, to keep the hype going.
12:52:16 <fizzie> (Disclaimer: guesswork and speculation.)
12:53:25 <oerjan> <Phantom_Hoover> AAGH the number 168 is stuck in my head and I don't know why. <-- it's your serial number, duh
12:54:22 <Phantom_Hoover> I strongly suspect it's an SCP, but I really don't want to find out which.
12:56:42 <fizzie> Few more bits of trivia that Knuth said about iTeX: no escape sequences, menu-driven, speech-recognition something. (From a random tweet.)
12:56:53 <fizzie> Heh, "knuth announcement: 27th most popular search in the past hour."
12:57:34 <fizzie> http://www.google.com/trends/hottrends?q=knuth+announcement
13:29:34 -!- pikhq has joined.
13:53:17 <Mathnerd314> I was thinking about XML in TeX, sometime in the past year... I'm pretty certain it could work (use <document> instead of \begin{document}, for example)
13:56:09 -!- hiato_ has joined.
13:57:04 -!- hiato has quit (Ping timeout: 260 seconds).
13:58:12 -!- hiato_ has quit (Client Quit).
13:58:20 -!- cpressey has joined.
13:58:32 -!- hiato has joined.
14:00:01 <cpressey> So I want to spider a site, but I also want to spider a version of it from a few years ago from www.archive.org.
14:00:22 <cpressey> I don't suppose a tool exists that does exactly that, so
14:00:43 <cpressey> Any suggestions for a spidering tool that is comfortably hackable?
14:05:46 <cpressey> wget -r plus some perl goo, it is, then.
14:07:44 <fizzie> Yes, there is already a XML syntax for TeX (TeXML). It's mostly intended for people who programmatically generate TeX code, though.
14:08:50 <fizzie> I don't have any clue how good/sensible it is.
14:10:09 <fizzie> It seems to do <env name="x"> instead of \begin{x} for example, so it's perhaps not that human-friendly to type directly.
14:16:03 -!- Quadrescence has quit (Ping timeout: 245 seconds).
14:21:55 -!- cpressey1 has joined.
14:22:33 -!- cpressey has quit (Quit: leaving).
14:22:47 -!- cpressey1 has changed nick to cpressey.
14:28:19 -!- MigoMipo has quit (Read error: Connection reset by peer).
14:29:10 -!- Quadrescence has joined.
14:32:59 <cpressey> This Knuth thing is an April Fool's joke, right? Just 3 months late?
14:37:27 <oerjan> presumably there's only one yearly tex conference...
14:37:53 <oerjan> top-level one, that is
14:41:13 -!- ais523 has joined.
14:48:31 <ais523> hmm, I proved both Reversible Brainfuck and DoFuck TC in my head last night
14:48:38 <ais523> although, as usual with proofs done mentally, there may be bugs
14:49:12 <cpressey> Indeed -- from the sound of it, Reversible Brainfuck would not be too hard to prove? Or is it weirder than its name suggests?
14:49:30 <ais523> (they're both BF with different definitions of the [ command; DoFuck has it enter the loop unconditionally, Reversible Brainfuck has it enter the loop if the current cell /is/ 0, rather than if it /isn't/ 0)
14:49:57 <cpressey> I was thinking I might actually implement my idea for Goldbach... was making some notes last night
14:50:03 <ais523> it's not trivial to prove; the issue is trying to make sure you can enter a loop the good-old-fashioned BF way
14:50:07 <cpressey> It will turn out to be very un-exciting, though
14:50:10 <ais523> which requires somehow ignoring the value when you enter it
14:52:21 <cpressey> My Goldbach idea just degrades into "Loop over pairs of primes until you find two that sum to the next larger even number you need", and if you can't do that, you can't perform a top-level loop. So if the Goldbach conjecture is false, there is a limit to the number of useful iterations you can do.
14:52:59 <cpressey> next larger even number -> first you have to goto -2, then -4, then -6 ... which all basicallly are gotos back to the first instruction of the program
14:53:07 <ais523> I remember asking bits of the mathematical community about the goldbach conjecture, though; they all believe it's true, just have no proof
14:53:26 <cpressey> I have a hard time fantasizing about it *not* being true.
14:53:46 <cpressey> Would there be just one "non-Goldbach" even number?
14:54:00 <cpressey> Nothing would seem to imply there would be more than one, if there were one.
14:54:01 <ais523> arguably 2, but that doesn't count
14:58:17 <Sgeo_> Am I supposed to fall in love with Lua?
14:58:47 <ais523> Sgeo_: it's good for some things, not for others
14:58:55 <cpressey> Heh... catseye.tc's front page looks pretty spectacular in FF when the canvas is only 23-or-so pixels high
14:59:15 * Sgeo_ is thinking of rewriting some of the C# stuff in Lua so it can be easily unloaded, changed, and loaded without restarting the bot
14:59:42 <cpressey> Sgeo_: Never a good idea to fall in love with a language. It can only lead to hurt when the summer's over.
15:00:25 <cpressey> Seriously, Lua's OK. In many respects, for me, it beats both Python and Ruby.
15:01:08 <cpressey> In others, well, nothing's perfect.
15:02:02 <Sgeo_> In what ways is Lua imperfect?
15:02:40 <cpressey> Er, well. There is no "standard" way to do object-orientation. So, if you mix and match two libraries which use different styles, it can be ugly.
15:03:05 <cpressey> I also something think they went too far, when they merged dictionaries and arrays into one thing.
15:04:12 <cpressey> Most of my other gripes are about the implementation and things like availability of libraries, which aren't really core language issues.
15:05:13 -!- Gracenotes has quit (Ping timeout: 264 seconds).
15:05:42 -!- Phantom_Hoover has quit (Ping timeout: 265 seconds).
15:05:53 -!- pikhq has quit (Read error: Connection reset by peer).
15:06:21 <cpressey> Where it beats Python: Lua doesn't make a distinction between attributes and dictionary entries. a['foo'] is the same as a.foo. The division in Python is artificial at this point, and in the code base I work on at least, extremely annoying.
15:06:55 <ais523> cpressey: JavaScript works like that too
15:07:07 <ais523> come to think of it, JS and Lua are surprisingly similar languages
15:07:26 <cpressey> Yes, JS is actually a lot better than I usually give it credit for. And yes, similar to Lua in many ways.
15:07:53 <ais523> JS has mostly been held back by being mostly trapped inside a web browser
15:08:35 <cpressey> Web browsers with generally lousy debugging/interaction capabilities, especially.
15:08:59 <cpressey> I think it's not a coincidence that much of the JS I write has the same feel as much of the Assembly I write.
15:09:07 <ais523> both Firebug and the Epiphany/Safari/Chrome web inspector are pretty good for debugging
15:09:31 <cpressey> Well yes, I didn't mean to imply such tools weren't available.
15:10:19 -!- relet has joined.
15:32:40 -!- pikhq has joined.
15:46:26 -!- hiato has quit (Quit: Lost terminal).
15:47:41 -!- hiato has joined.
15:50:45 -!- hiato has quit (Client Quit).
15:51:08 -!- Quadrescence has quit (Ping timeout: 265 seconds).
15:52:47 -!- hiato has joined.
15:55:29 -!- Quadrescence has joined.
16:05:10 -!- KingOfKarlsruhe has joined.
16:06:32 -!- ais523 has quit (Read error: Operation timed out).
16:06:57 -!- ais523 has joined.
16:25:56 -!- Quadrescence has quit (Ping timeout: 265 seconds).
16:26:35 -!- hiato has quit (Quit: underflow).
16:27:09 -!- ais523 has quit (Ping timeout: 240 seconds).
16:30:30 -!- hiato has joined.
16:30:49 -!- hiato has quit (Client Quit).
16:31:15 -!- hiato has joined.
16:31:19 -!- Quadrescence has joined.
16:31:22 -!- hiato has quit (Client Quit).
16:34:06 -!- hiato has joined.
16:38:09 -!- augur has quit (Remote host closed the connection).
16:42:07 -!- pikhq has quit (Quit: New kernel; back in a bit.).
16:42:44 -!- KingOfKarlsruhe has quit (Remote host closed the connection).
16:43:27 -!- kar8nga has joined.
16:45:02 -!- pikhq has joined.
16:59:25 -!- oerjan has quit (Quit: leaving).
17:04:02 -!- kar8nga has quit (Remote host closed the connection).
17:11:46 -!- hiato has quit (Quit: underflow).
17:12:27 -!- hiato has joined.
17:13:37 -!- hiato has quit (Client Quit).
17:14:10 -!- hiato has joined.
17:22:58 -!- KINGZ has joined.
17:39:03 -!- KINGZ has left (?).
17:43:18 -!- ais523 has joined.
18:01:38 -!- Quadrescence has quit (Ping timeout: 265 seconds).
18:08:05 -!- augur has joined.
18:23:40 -!- ipatrol has joined.
18:23:52 -!- ipatrol has left (?).
18:30:24 -!- kar8nga has joined.
18:30:26 -!- kar8nga has quit (Read error: Connection reset by peer).
18:40:47 -!- Geekthras has quit (Ping timeout: 265 seconds).
18:53:23 -!- Geekthras has joined.
19:05:30 -!- coppro has joined.
19:10:13 -!- BeholdMyGlory has quit (Ping timeout: 260 seconds).
19:10:38 -!- Gracenotes has joined.
19:17:34 -!- Quadrescence has joined.
19:22:52 -!- ais523 has quit (Remote host closed the connection).
19:23:14 -!- ais523 has joined.
19:26:22 -!- Geekthras has quit (Ping timeout: 276 seconds).
19:34:25 -!- Geekthras has joined.
19:45:46 -!- coppro has quit (Quit: Reconnecting…).
19:47:41 -!- Gracenotes has quit (Ping timeout: 276 seconds).
19:51:09 -!- poiuy_qwert has joined.
19:54:09 -!- coppro has joined.
19:57:21 -!- ghostwriter42 has joined.
19:58:16 <ghostwriter42> quick i need a someone to pick a *even* number between 50 and 100 that has two different digits!
20:00:41 -!- ghostwriter42 has left (?).
20:00:47 -!- ghostwriter42 has joined.
20:00:59 <Sgeo_> Is 68 a common response or something?
20:01:29 <ghostwriter42> http://mindcontrol101.blogspot.com/ read the paragraph that says "pick a number"
20:03:04 -!- ghostwriter42 has left (?).
20:07:10 -!- coppro has quit (Ping timeout: 252 seconds).
20:07:32 -!- pikhq has quit (Read error: Connection reset by peer).
20:11:21 -!- coppro has joined.
20:15:21 -!- Geekthras has quit (Ping timeout: 240 seconds).
20:16:23 -!- pikhq has joined.
20:18:02 -!- impomatic has joined.
20:20:13 <impomatic> Can anyone think of a cool name for a website specialising in programming games?
20:21:13 <coppro> no, but I am interested
20:21:45 -!- BeholdMyGlory has joined.
20:21:45 -!- BeholdMyGlory has quit (Changing host).
20:21:45 -!- BeholdMyGlory has joined.
20:24:13 <impomatic> It's a shame programming.com, programming.net and programming.co.uk are wasted :-(
20:28:55 <ais523> impomatic: btw, there was a bit of movement on the BF Joust leaderboard recently
20:28:57 <ais523> so it isn't /quite/ dead
20:32:38 * pikhq can has decent headphones! And stuff for curry, reubens, and such!
20:33:08 -!- Gracenotes has joined.
20:35:34 -!- Gracenotes has quit (Excess Flood).
20:38:34 -!- CakeProphet has joined.
20:52:20 -!- MizardX- has joined.
20:56:43 -!- MizardX has quit (Ping timeout: 276 seconds).
20:56:45 -!- MizardX- has changed nick to MizardX.
20:56:52 -!- Behold has joined.
20:57:09 -!- BeholdMyGlory has quit (Disconnected by services).
20:57:14 -!- Behold has changed nick to BeholdMyGlory.
21:00:31 -!- Gracenotes has joined.
21:04:51 -!- Geekthras has joined.
21:21:21 -!- Geekthras has quit (Ping timeout: 260 seconds).
21:36:34 -!- impomatic has quit (Quit: ChatZilla 0.9.86 [Firefox 3.5.10/20100504093643]).
22:06:55 -!- coppro has quit (Remote host closed the connection).
22:39:57 -!- ehirdiphone has joined.
22:40:35 <ais523> good point, I haven't heard any followup on that
22:40:40 <ais523> I don't know, is the answer
22:40:47 <ais523> which surprises me, I thought it would have at least hit Slashdot
22:41:32 <ais523> just checked a search: it seems he was joking
22:41:45 <ais523> he announced an XML-based TeX, called iTex
22:41:47 <pikhq> ehirdiphone: It was a joke.
22:41:53 <ehirdiphone> ais523: Guess (stolen from Hacker News): a typo was found in a comment in METAFONT
22:42:29 <pikhq> iTeX features Unicode, XML syntax, 3D printing, stereophonic sound, and a menu-driven interface.
22:42:36 <ais523> only places the actual announcement seems to have hit are Twitter and ycombinator.com
22:42:54 <ehirdiphone> Yeah, Unicode in TeX. Hahaha, what a silly idea.
22:43:40 <ehirdiphone> Remind me to eat after leaving. I'm starving.
22:43:56 <ehirdiphone> Have to scavenge something from my room...
22:45:03 <ehirdiphone> pikhq: Are filesystem drivers servers in HURD?
22:46:25 <pikhq> ehirdiphone: Quite a few hardware drivers.
22:46:55 <pikhq> Mouse, keyboard, text console.
22:47:17 <pikhq> I think also the block devices.
22:47:26 <pikhq> No, it doesn't have USB.
22:47:39 <pikhq> Because Linux 2.2 didn't, and they use Linux 2.2 drivers.
22:48:11 <pikhq> Would take approximately 0 work to make them be in userspace.
22:48:31 <pikhq> As you can make a server for literally ANY FILE. :)
22:49:55 <pikhq> No, the kernel does not even acknowledge the existence of a filesystem.
22:50:36 <pikhq> Mach provides message passing.
22:51:13 <ais523> if it's unaware of the existence of filesystems, is it technically a kernel?
22:51:20 <pikhq> ehirdiphone: The Mach microkernel that runs in kernelspace.
22:51:50 <pikhq> ehirdiphone: Then it's hard to define what the kernel is. Everything else is daemons.
22:53:06 <ehirdiphone> pikhq: Plan X will have a concept of / in the kernel, but only because the *whole OS* will be based on per-process namespaces.
22:53:23 <ais523> the kernel's / is a different / from everyone else's?
22:53:46 <ehirdiphone> ais523: In plan 9, every process has a different /
22:54:01 <pikhq> Whereas in HURD, a "file" is nothing more than a name provided for a port of a server.
22:54:07 <ais523> there's a global / too, though
22:54:11 <ais523> so the things can actually communicate
22:54:24 <pikhq> And most of the actual POSIX abstractions are coming out of libhurd.
22:54:28 <ehirdiphone> ais523: E.g. The rio wm works by rebinding the /dev/screen of its children, for instance.
22:54:56 <ehirdiphone> ais523: Well, process 1's namespace but thats irrelevant.
22:55:17 <ehirdiphone> They communicats by inheriting files served by other processes.
22:55:30 <ais523> so if you reference, say, /home/alise, how does Plan 9 resolve that?
22:55:42 <cpressey> The kernel's the part that, when it crashes, you're fucked.
22:55:59 <pikhq> cpressey: Then that's Mach.
22:56:10 <ais523> ehirdiphone: right, I forgot how Plan 9 actively hates the FHS
22:56:13 <ais523> rather than just ignoring it
22:56:18 <ais523> cpressey: that's a good definition
22:56:19 <pikhq> And *maybe* the authentication daemon.
22:56:31 <AnMaster> cpressey, only works if you have an MMU and such
22:56:37 <ehirdiphone> ais523: Your shell's /usr/alise. The same as your login process's, prolly.
22:56:44 <pikhq> But really, that's only if you happen to not be running a subHurd.
22:56:49 <AnMaster> cpressey, on classic MacOS, any program could result in that
22:57:06 <pikhq> (one can run a virtual Hurd system by starting another instance of the daemons.)
22:58:00 <pikhq> ais523: Plan 9 also has absolutely no reason to acknowledge the FHS. It is very much not Unix.
22:58:26 <ehirdiphone> pikhq: Well. It's the official successor to 10th Edition Unix.
22:58:41 -!- Mathnerd314_ has joined.
22:59:10 <ehirdiphone> ais523: /usr for users dates back to PDP unix.
22:59:15 <pikhq> ehirdiphone: Yes, and Windows is the official successor to DOS, but that doesn't make it DOS.
22:59:20 -!- nooga has joined.
22:59:25 -!- Mathnerd314 has quit (Disconnected by services).
22:59:40 <ehirdiphone> ais523: You should know; you tried them out.
22:59:50 <nooga> are there some 'reverse parsers' for generating random expressions that match specified syntax?
22:59:57 -!- Mathnerd314_ has changed nick to Mathnerd314.
23:00:06 <nooga> because i wrote one and i don't know if it's worth releasing
23:00:15 -!- ehirdiphone has changed nick to alisephone.
23:00:43 <nooga> one person is enough :D
23:01:01 <ais523> I wrote one for ICA a whie back because I needed testcases
23:01:02 <nooga> don't know, maybe it's not me :D
23:01:03 <ais523> but it didn't work very well
23:01:38 <alisephone> ais523: Read "...because I needed testicles." :|
23:01:56 <ais523> nah, I have those already
23:02:31 <alisephone> pikhq: I plan to not have virtual consoles in kernel.
23:02:55 <ais523> idealized concurrent algol, it's one of the languages I work with in my day job
23:03:03 <alisephone> They'll be servers that hook into the video and keyboard daemons or something.
23:03:14 <pikhq> alisephone: HURD doesn't either.
23:03:23 <ais523> it's... basically, algol designed to work better for mathematicians
23:03:25 <pikhq> It has the *physical* console in kernel.
23:03:36 <ais523> as in, not to be easier to use, but to be easier to analyze mathematically
23:03:36 <pikhq> Because you kinda need to be in kernel-space to write to the VGA buffer.
23:03:52 <ais523> it's what computer scientists use if they want a vaguely imperative language to work with
23:04:08 <pikhq> ... How do you think a non-framebuffer text console works?
23:04:48 <pikhq> It writes ASCII and color info interspersed to the VGA buffer. Glee.
23:05:47 <alisephone> pikhq: Ooh, in true Plan tradition I get to improve C.
23:06:45 <nooga> alisephone: still hacking Plan9?
23:07:03 <alisephone> (A,B,C) = strict { A [0]; B [1]; C [2]; }
23:09:05 <ais523> <jk1150> Even in the most recent version of Lotus Notes, you will get an error if you are typing in the subject line and click the add attachment button. IBM says this is a feature because you cannot add an attachment to a subject line, I think it is disgusting.
23:09:08 <nooga> my crappy gsoc proposal for P9 was not accepted so i left the topic for a while
23:09:18 <ais523> I like the reasoning...
23:09:30 <nooga> now i'm playing with rails 3 & heroku
23:09:33 <alisephone> In fact, let it deconstruct arbitrary structs
23:10:44 <nooga> webdev is my current job
23:11:04 <nooga> and i do a lot of PHP which @#(&(&(@#**(#(*@#*#@*@# SUCKS
23:11:05 <alisephone> pikhq: Oh, and perhaps actual extensible types?
23:11:47 <nooga> i couldn't even find non-irritating MVC stack for PHP so i tried to write my own and failed :D
23:12:15 <alisephone> pikhq: Does Hurd have a concept of users?
23:12:39 <pikhq> alisephone: That's done by the authentication daemon.
23:12:46 <alisephone> In Plan 9ish systems, users are... A login process.
23:13:15 <ais523> in other news, this is my first nday playing B Nomic
23:13:34 <ais523> Teucer challenged my claim to be a newbie, and I was as shocked as he was
23:13:41 <ais523> strange the way the rules work out, sometimes
23:13:49 <alisephone> You don't seriously buy the Era 4-5 argument?
23:13:50 <ais523> you really can't take much about B's history for granted
23:14:01 <ais523> alisephone: eras 4 and 5 never happened, BGora didn't have ndays
23:14:12 <ais523> and yes, most of B buys that argument
23:14:38 <ais523> platonically, I mean; sure, people played through and enjoyed them, and we can still talk about events happening in them
23:14:42 <ais523> but they were ignored by the rules
23:14:53 <ais523> to be precise, they happened but the Clock was off continuously
23:15:04 <alisephone> I distinctly remember the argument sucking.
23:15:37 <alisephone> I think my nomic school is "formalism". :)
23:16:36 <alisephone> Alma mater: RMSN (The Retarded Monkey School of Nomic).
23:16:46 <ais523> alisephone: the rule allowing the clock to be switched back on was commented out
23:17:05 <ais523> and even if the original comment crisis argument sucked slightly, someone made a much better one later on that pointed out it was commented out with both definitions
23:18:14 <alisephone> ais523: Pretty sure B has never existed :P
23:18:17 <nooga> my flatmates are using too much bandwith
23:18:47 <nooga> some idiot secured the router
23:18:58 <ais523> alisephone: heh, I'm playing B under the name "703B E29B E9CC E4ED A7E2 7F62 1608 627B 1BA5 7726" because the requirement to be uniquely named is back, and I feared "ais523" might not be unique enough
23:19:49 -!- FireFly has joined.
23:20:25 <ais523> meh, it's a GPG key fingerprint
23:20:30 <ais523> I'm pretty sure it's unique
23:20:56 <ais523> there are a suspiciously large number of Es in the first half, too
23:22:42 <ais523> there's quite a difference
23:22:58 -!- Gregor-W has joined.
23:23:44 <ais523> Warrigal: an API to allow esoprograms to use operating system services, by Sgeo
23:23:47 <Gregor-W> PSOX is a system interface utilizing only stdout and stdin, so that it can be used by very restricted (esoteric) programming languages without modification.
23:23:54 <ais523> but its design made it basically only usable with BF
23:24:12 <ais523> because it relied on literal NUL characters a lot
23:24:19 <ais523> oh, I suppose Befunge can do those too
23:24:24 <ais523> but many esolangs can't
23:25:37 <Warrigal> I hope it was a binding to the C standard library.
23:25:58 -!- nooga_ has joined.
23:26:08 <Gregor-W> Of course. The only magic to it was that the interface between the language and it was very thin, just stdout/stderr.
23:26:22 <Gregor-W> It itself was just linked however it works on the host.
23:26:59 <ais523> meanwhile, reddit are debating the "M-x google-maps" command
23:27:07 -!- nooga has quit (Ping timeout: 260 seconds).
23:29:12 -!- tombom has quit (Quit: Leaving).
23:33:31 <ais523> AnMaster: not by default, someone implemented it
23:33:51 <ais523> I /hope/ it'll never get into the standard distribution
23:33:52 <AnMaster> ais523, what does it do? use picture-mode to display images?
23:34:03 <ais523> I think it's more complex than that, I haven't looked into it
23:34:32 <ais523> http://julien.danjou.info/blog/2010.html#M%2Dx%20google%2Dmaps
23:37:38 <ais523> as I said, I don't know the emails
23:37:41 <ais523> *don't know the details
23:37:48 <ais523> but Emacs is certainly capable of showing images
23:56:39 -!- FireFly has quit (Quit: ClIRC - IRC client for Nintendo DS).
23:56:42 <alisephone> Warrigal: No. Sgeo invented his own APIs.
23:58:44 <Sgeo_> Besides the NUL thing, howso?
23:59:04 <ais523> Knuth "also stated that this successor of TeX will have features like 3-D printing, animation, stereographic sound."
23:59:10 <ais523> hmm, has hit Slashdot, eventually