00:04:19 -!- Slereah has joined.
00:04:19 -!- Slereah_ has quit (Read error: 104 (Connection reset by peer)).
00:16:43 -!- Linus` has changed nick to Linus.
00:28:25 -!- Linus has quit ("Puzzi. Sì, parlo proprio con te. Puzzi.").
00:42:13 -!- sebbu has quit ("@+").
00:52:27 -!- oerjan has quit ("Good night").
00:59:00 -!- Slereah has quit (Read error: 113 (No route to host)).
01:20:29 -!- Slereah_ has joined.
02:01:10 <GregorR> http://www.crazymonkeygames.com/fullscreen.php?game=Pandemic-2 // this game is far more fun than it should be given that the goal is to exterminate humanity.
02:03:51 <Slereah_> Pandemic is actually ^pretty meh
02:04:02 <Slereah_> Once you get how it works, there's no challenge
02:19:47 -!- cathyal has quit.
02:32:08 -!- Judofyr has quit (Read error: 104 (Connection reset by peer)).
02:32:44 -!- Judofyr has joined.
02:41:59 -!- lament has joined.
02:42:12 <optbot> lament: is amount of coffee consumed actually related to amount of sleep?
02:51:27 <psygnisfive> interestingly, there have been some ideas about how to most effectively waken up with coffee when you're getting sleepy
02:51:41 <psygnisfive> drink a strong cup of coffee and nap for about 15 minutes to half an hour
02:51:50 <psygnisfive> by the time you wake up, the caffeine will be kicking in in full force
02:52:12 <psygnisfive> and having napped, you've gotten some sleep thus reducing your sleepiness
02:56:43 <psygnisfive> so you me and oklopol could have a threesome. :D
03:01:14 <bsmntbombdood> http://www.ectomo.com/wp-content/uploads/2007/08/footpussy.jpg
03:07:08 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | I don't name my software like that.
03:07:57 * zbrown stabs all operating systems
03:10:46 <lament> psygnisfive: do you drink coffee with sugar?
03:11:22 <lament> sugar kicks in very quickly
03:21:40 <bsmntbombdood> especially delicious when extracted under high pressure
04:07:32 -!- immibis has joined.
04:29:40 <GregorR> It's time to play "spot the GIMPing"!
04:29:46 <GregorR> http://codu.org/pics/other/pec2.jpg // spot the GIMPing!
04:31:50 -!- Sgeo has joined.
04:42:43 <Asztal> nice ball, but the architecture is obvious
04:53:26 <Asztal> hmm, http://codu.org/hats/BritDrivingCap-sm.jpg is a flat cap? I didn't recognise it from that angle, thought it was something else :)
04:54:03 <GregorR> Reload pec2.jpg , I made some fixes.
04:55:22 <Asztal> Looks a lot better now. I think needs some sort of reflection in the cylindrical shiny thing, though, even if it is blurry. And maybe a bit of a shadow.
04:56:40 <GregorR> I don't think it should be reflected ...?
04:56:50 <GregorR> (That is, I don't think my feet are visible)
04:57:45 <GregorR> Please do, that's the skill I lack :P
04:58:16 <psygnisfive> one: the shadow on the ball is dark but the grating isnt as dark
04:58:49 <psygnisfive> look at the direction the light is coming from on your body
04:58:55 <psygnisfive> its coming from the left, reflecting off the wall
04:59:22 <psygnisfive> you're standing in the shadow of a pillar. wheres that light coming from on the ball?
04:59:58 <GregorR> The ball should be lit purely by ambient light I suppose.
05:00:25 <psygnisfive> the lights just coming from the wrong direction
05:00:37 <GregorR> Where should it be coming from? Up and left?
05:00:44 <GregorR> That's what I had before and it seemed funky to me.
05:01:26 <psygnisfive> with some minor specs on the far left and right sides like your shoes
05:09:00 <GregorR> psygnisfive: Refresh. Better or worse?
05:09:52 <psygnisfive> better with the light, but the shadows are still off.
05:10:04 <psygnisfive> look around your feet, see how the shadows on the grating are?
05:10:18 <GregorR> I see how they are, I have no idea how to replicate that.
05:11:06 <GregorR> I have to do this manually, don't I X-P
05:14:23 <oklopol> for absolute photorealism, i recommend filling the screen with #000000
05:14:24 <GregorR> Yeah, I know what you're saying, it's just a bit more tedious than "add a dark area and fade it to 50%" :P
05:15:24 <psygnisfive> just hand paint a new layer with some black
05:16:10 <GregorR> Yeah, but hand painting = a process X-P
05:16:12 <oklopol> yes, just add a layer of black jesus christ
05:17:29 <GregorR> Everything is fine with some black Jesus Christ.
05:19:37 <GregorR> Of course, this time it's uploading SUUUPERSLOOOOOOOWWWWWWWLLLLYYYY
05:22:44 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
05:23:05 <GregorR> psygnisfive: Rererererererelook for me? :)
05:23:55 <psygnisfive> better. make it darker right where the ball meets the grate
05:24:29 <oklopol> GregorR: your swing looks pretty fake too
05:24:47 <GregorR> oklopol: That was an actual swing -_-
05:24:58 <oklopol> GregorR: then maybe i'm calling you a nerd :P
05:25:07 <GregorR> oklopol: But it's difficult to swing well when you're being SOAKED IN EFFING FREEZING COLD WATER
05:26:42 <oklopol> yes, GregorR, you should hold mallots like cocks, not like pussies
05:26:51 <oklopol> because, you know, they're sticks
05:27:05 <oklopol> psygnisfive: indeed i didn't
05:29:08 <psygnisfive> infact, bsmntbombdood, you, me, and your girl friend
05:29:17 <oklopol> if you really want me to answer, i guess i could decline, but i'm not sure if you'll prefer that :D
05:30:23 <oklopol> perhaps #esoteric should makes the worlds largest marriage graph
05:31:27 <oklopol> see you later, math exam fun ->
05:31:50 <bsmntbombdood> you don't them them mormons have a dude with more than 36 wives?
05:32:04 <GregorR> That was quite the sentence.
05:32:44 <bsmntbombdood> *you don't think those mormons have a dude with more than 36 wives?
05:33:41 * psygnisfive pulls out two equally awesome ring algebras
05:35:01 <omniscient_idiot> bsmntbombdood: I don't think many mormoms are married to irc bots.
05:35:19 * GregorR swaps out his personality for a bit of fun.
05:35:26 <bsmntbombdood> i've always been interested in teledildonics platforms...
05:36:03 <bsmntbombdood> oh by the way, does anyone here have access to springerlink through their uni?
05:50:26 -!- oklopol has quit (Read error: 110 (Connection timed out)).
05:59:02 -!- Judofyr has quit (Read error: 104 (Connection reset by peer)).
05:59:35 -!- Judofyr has joined.
06:27:10 -!- immibis has quit ("Hi Im a qit msg virus. Pls rplce ur old qit msg wit tis 1 & hlp me tk ovr th wrld of IRC. and dlte ur files. and email ths to).
06:56:10 <fizzie> The fungot crash log looks like it has generated some babble without the correct terminating '2' in it.
06:56:10 <fungot> fizzie: lemme see if i can
06:58:48 <fizzie> I'm still not sure why in a stack "0 -29 -25 -21 ... 87 91 95 46947 3" executing STRN's 'P' instruction turns the stack into "... 87 91 95 99 3".
07:05:32 -!- immibis has joined.
07:06:25 <immibis> i should find a hobby besides annoying people with irc bots
07:06:50 <fizzie> Is that "using bots to annoy people" or just "annoy people who have bots"?
07:06:59 <immibis> using bots to annoy people
07:07:01 -!- CoffeeBot has joined.
07:07:08 -!- CoffeeBot has quit (Read error: 54 (Connection reset by peer)).
07:07:53 <fizzie> I guess it could've also been that "people who are both annoying and have bots" was the hobby.
07:08:08 <immibis> no because that doesn't make sense
07:12:59 <Jiminy_Cricket> To the general population, irc bots themselves don't make sense.
07:17:03 <fizzie> My bot doesn't make any sort of sense either. There's no way it could end up in the text output place without having a terminating 2 on the output row. There must be some other issue.
07:35:32 -!- toBogE has joined.
07:44:15 -!- CoffeeBot has joined.
07:44:15 <optbot> CoffeeBot: just don't claim I made it :)
07:44:16 <optbot> fungot: running qbf results in some state |S>
07:44:16 <fungot> optbot: i have that down, i'll have money
07:44:16 <optbot> fungot: what are the threads?
07:44:16 <fungot> optbot: huh? only tests a single bit index for pheromones are not fnord
07:44:17 <fungot> optbot: ummm...i dont know. that is
07:44:18 <optbot> fungot: this isn't branfuck?
07:44:18 <fungot> optbot: the latter. see the topic)
07:44:36 -!- CoffeeBot has left (?).
07:44:48 <immibis> bug in optbot: what is branfuck?
07:45:03 <immibis> oh, whatever. In fatbot then
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:00:44 -!- olsner has quit ("Leaving").
08:03:25 -!- fungot has quit ("just testing...").
08:04:25 -!- fungot has joined.
08:06:28 <fizzie> Haven't fungotized the Befungized Underload interp yet, but I bumped the cycle count up a bit.
08:06:28 <fungot> fizzie: they are talking to you
08:08:14 <fungot> ^<lang> <code>; ^def <command> <lang> <code>; ^show [command]; lang=bf, code=text/str:N; ^str 0-9 get/set/add [text]
08:08:42 <fungot> <CTCP>.. !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€‚ƒ„…†‡ˆ‰Š‹ŒŽ‘’“”•–—˜™š›œžŸ ¡¢£¤¥¦§¨©ª«¬®¯°±²³´µ¶·¸¹º»¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏ ...
08:11:24 <fizzie> Newlines are filtered a bit.
08:11:32 <fizzie> But that Underload interp is the brainfuck one.
08:11:34 <fungot> >,[>,]<[<]>[<+4[>-8<-]+>-[-7[-2[<+3[>-4<-]+>[<+4[>-5<-]+>[-11[-3[[-]<2[>[-]>+<2-]>>[<2+>>-]+<[->-<3[[>+<-]<]>>[>]]>[->[>]<[[>+<-]<]<2[[>+<-]<]<[[>+<-]<]>>[>]>[[[>]>+<2[<]>-]<2[[>+<-]<]>>[>]>[>]>[<2[<]<[<]<+>>[>]>[>]>-]<2[<]>]>>[[<+>-]>]<2[<]]]<[->>[>]<[[>>+<2-]<]<2[[>+<-]<]>+>[>]+5[>+8<-]+2>-[<+[<]>+[>]<-]]>]<[->>[[<2+>>-]>]<3[[>+<-]<]]>]<[-<[[<]>.[-]>[[<+>-]>]>>[[<+>-]>]<2[<]<2]>>>[[<+>-]>]<2[<]<]>]<[->>[>]<[[>+<-]<]<2[>>>>[>]
08:11:51 <fizzie> It used to be too slow to run that (:aSS):aSS quine.
08:15:49 <AnMaster> Deewiant, I ran some coverage analysis on efunge when running mycology, you never test wrapping straight up or straight down (so the cardinal y wrapping code is never hit)
08:16:38 <AnMaster> Deewiant, nor do you go out of bounds above or below the code at all
08:17:08 <fizzie> AnMaster: It might interest you to know that fungot's now running on cfunge. (Testing whether the occasional hangup might be a RC/Funge bug. Probably isn't, but you never know.)
08:17:09 <fungot> fizzie: cat /dev/ mem? how does one convert ( ' ( n e v e r s e))
08:17:43 <AnMaster> fizzie, well you may hit some cfunge bugs, if you do, report
08:18:29 <AnMaster> fizzie, also I got no idea if rc/funge is valgrind clean, but a debug build of cfunge is, apart from not freeing the handle list used in SOCK and FILE. (So two "still reachable")
08:18:38 <AnMaster> a release build will give a lot more still reachable
08:19:21 <AnMaster> Deewiant, also you don't test ) for negative values of count it seems :)
08:19:25 <Deewiant> I assume that if you can get it to work in one direction you can get it to work in any other direction.
08:19:54 <AnMaster> indeed, I just had breakfast. Still sleepy...
08:20:14 <Deewiant> hmm, it /should/ try ) with a negative count
08:20:26 -!- oklocod has joined.
08:20:30 <Deewiant> maybe I typoed and it does ( twice
08:21:10 <fizzie> What should () with negative count do?
08:21:33 <fizzie> But is there some sort of sensible behaviour?
08:21:38 <AnMaster> Deewiant, well coverage analysis claims you don't try ) with negative
08:21:57 <Deewiant> AnMaster: but does mycology output "UNDEF: ) with a negative"...
08:21:59 <AnMaster> fizzie, 1) reflect (what I do) 2) pop |count| 3) other
08:22:08 <AnMaster> UNDEF: ( with a negative count reflects and pops 0 times or less than the absolute value of the count
08:22:08 <AnMaster> UNDEF: ) with a negative count reflects and pops 0 times or less than the absolute value of the count
08:22:33 <Deewiant> ah well, I guess I'll have to fix that then
08:23:56 <fizzie> Hacked in that "chroot+setuid after starting" thing so I don't have to have to bother with a real chroot with libraries and everything.
08:23:58 <AnMaster> wtf did someone mess with some bot to make it /msg me?
08:24:18 <AnMaster> "* CoffeeBot__ is making a coffee in an office mug with cold milk to help him wake up for you"
08:24:21 <immibis> i asked it to make coffee on account of you being asleep
08:24:53 <AnMaster> immibis, I don't drink coffee, this morning I had some fruit juice, and a slice of bread
08:27:13 <AnMaster> Deewiant, you never attempts to unload a valid fingerprint when none is loaded? I'm unsure, it may be that the case is detected earlier in my code than in the fingerprint instr stack popping code.
08:28:11 <AnMaster> no I can't see any obvious place where it would have been detected...
08:29:38 <immibis> oh anmaster your away is set to sleeping
08:30:02 <AnMaster> Deewiant, nor do you test MODU's signed division for positive x and y I think
08:30:22 <Deewiant> probably because it's not interesting
08:30:35 <Deewiant> again, if it can get it right for negatives it's probably right for positives.
08:31:14 <AnMaster> FIXP's rand() for negative arguments (or equal to 0) isn't tested either
08:31:23 <AnMaster> and that could be worth testing
08:31:31 <AnMaster> if the implementation does rand() % argument
08:32:12 <AnMaster> FIXP's acos() never ends up hitting inf or nan either.
08:32:36 <AnMaster> it's sqrt() doesn't seem to be tested on negative numbers
08:33:15 <AnMaster> CPLI div isn't tested for cases when Bi * Bi + Br * Br == 0
08:34:39 <AnMaster> oh and CPLI's abs() doesn't hit nan or inf either, but that may not be possible. (Unsure).
08:35:19 <Deewiant> there are edge cases everywhere, I can't be bothered to test every single one.
08:36:05 <AnMaster> Deewiant, the FIXP randomness one could be worth testing, since I'm sure at least cfunge would have gotten division by zero there originally. (It doesn't since a few months, due to my fuzz testing)
08:38:48 <AnMaster> immibis, glad you didn't forget to remove the milk from the orange juice :P
08:41:02 <immibis> theres currently three ghosts of coffeebot online due to lag and apparently the server is not disconnecting them
08:41:30 <immibis> ...ok the first disconnected
08:42:18 <AnMaster> immibis, I got two messages from that bot both times
08:42:45 <AnMaster> so two of them at least were alive at the same time
08:43:25 <immibis> no i sent the first, it lagged and got ghosted, so i restarted it and sent the second
08:43:46 <immibis> its a very badly coded bot. it needs recompiling to change the nick it connects with, and if the nick is in use it does nothing
08:45:16 -!- CoffeeBot has joined.
08:45:43 <oklocod> immibis: easy way out, add a "raw" command
08:45:59 <ais523> immibis: ThutuBot needs recompiling to change its nick too...
08:46:06 <AnMaster> immibis, as it is here is it 1) coded in an esolang or 2) runs esolang related stuff?
08:46:08 <oklocod> oh, well doesn't help if nick is in use to begin with
08:46:15 <AnMaster> ais523, didn't know you were here
08:46:20 -!- CoffeeBot has left (?).
08:46:31 <ais523> so I'll be uncommunicative for a while
08:46:57 -!- oklocod has changed nick to oklopol.
08:47:07 <fizzie> fungot needs source-code-changes to change the nick too (surprise!) and also doesn't really handle the nick-in-use case.
08:47:08 <fungot> fizzie: " yes." however, i cannot put a let statment are evaluated is compiler-defined.
08:47:22 <fizzie> fungot: That's not a real explanation.
08:47:22 <fungot> fizzie: i must depart. but, yeah, i don't really know
08:47:42 <AnMaster> oklopol, ... It just sounded funny in Swedish kind of
08:48:38 <oklopol> AnMaster: i was half making fun of you, and half actually laughing at the typo, so neither of you need be offended!
08:48:39 <AnMaster> so in my head it turned out somewhat like ionized by mixing the Swedish meaning with the English -ed suffix.
08:48:51 <fizzie> oklopol: Are you sure that doesn't mean both can get offended.
08:50:13 <AnMaster> fizzie, those changes for chroot + setuid means you need to start it as root right?
08:50:27 <AnMaster> hm... I wonder how that interacts with resource limits...
08:50:29 <ais523> most OSs only let root chroot things
08:50:46 <AnMaster> ais523, he could make the binary suid root ;P
08:50:57 <AnMaster> and also how it interacts with PERL.
08:51:07 <ais523> where non-root can chroot things IIRC
08:51:20 <fizzie> Yes. (Well, you could do it with the correct capabilities too, I guess.) It drops root privileges immediately after parsing the getopt results, though.
08:51:24 <AnMaster> ais523, really? Well I doubt it means anything for security in cygwin.
08:51:25 <ais523> although it doesn't really work like a typical secure chroot, as all the Win32 API functions just ignore it
08:51:35 <ais523> AnMaster: yep, it's just there so chroot stuff works
08:52:11 <AnMaster> fizzie, hm if it drops it there then it needs to have the source file in the chroot
08:52:43 <fizzie> Although the user it setuid()s to only has read permissions for the source file, not write.
08:54:04 <AnMaster> fizzie, fun fact: avoid loading your language model data using i if you ever planned changing, cfunge uses mmap() to simply reading (handling \r\n across the boundary between two fread() chunks was just too painful...)
08:55:59 <fizzie> I don't think I'll ever try to get that language model stuff to the funge-space; it seems to work just fine by simply seeking around the file and reading few bytes here and there.
08:56:57 <fizzie> cfunge's quite a bit faster than RC/Funge, though. Changed the brainfuck interpreter amount-of-instructions limit from 200k to 600k, and it still says "out of time" in a reasonable time, I think
08:57:29 <AnMaster> and yes cfunge is fast, that was one of the design goals
08:57:36 <AnMaster> tracing slows it down a lot though
08:57:39 <fizzie> And 600k is enough for the Underload quine, most importantly.
08:58:06 <ais523> it worries me that a quine would need a limit of 600k...
08:58:09 <AnMaster> since it traces to stderr, and stderr is unbuffered
08:58:41 <AnMaster> fizzie, 600 * 1024 bf instructions?
08:58:54 <ais523> (the second one is Keymaker's version, it works a different way)
08:59:11 <fizzie> AnMaster: Actually 600000. And it's bytecode instructions, so things like +++++ are a single instruction.
08:59:25 <fungot> >,[>,]<[<]>[<+4[>-8<-]+>-[-7[-2[<+3[>-4<-]+>[<+4[>-5<-]+>[-11[-3[[-]<2[>[-]>+<2-]>>[<2+>>-]+<[->-<3[[>+<-]<]>>[>]]>[->[>]<[[>+<-]<]<2[[>+<-]<]<[[>+<-]<]>>[>]>[[[>]>+<2[<]>-]<2[[>+<-]<]>>[>]>[>]>[<2[<]<[<]<+>>[>]>[>]>-]<2[<]>]>>[[<+>-]>]<2[<]]]<[->>[>]<[[>>+<2-]<]<2[[>+<-]<]>+>[>]+5[>+8<-]+2>-[<+[<]>+[>]<-]]>]<[->>[[<2+>>-]>]<3[[>+<-]<]]>]<[-<[[<]>.[-]>[[<+>-]>]>>[[<+>-]>]<2[<]<2]>>>[[<+>-]>]<2[<]<]>]<[->>[>]<[[>+<-]<]<2[>>>>[>]
09:00:29 <AnMaster> fizzie, oh an idea, [-] into c (clear cell)
09:00:34 <fizzie> Around 800 instructions unless I missed some other instruction in-between.
09:00:35 <fungot> echo reverb rev bf rot13 hi rev2 fib wc ul ctcp
09:00:46 <fizzie> No, ctcp is immediately after ul; so 800 instructions, then.
09:01:15 <fizzie> They're in the state file in the same order as in ^show.
09:01:29 <fizzie> I looked into that to see how long the program was when compiled into that bytecode.
09:02:12 <AnMaster> oh and G in FILE is quite ineffective in cfunge.
09:02:20 <AnMaster> so avoid it in performance critical code
09:02:30 <ais523> AnMaster: you mean "inefficient", I think
09:02:34 <ais523> "ineffective" means "doesn't work"
09:02:44 <AnMaster> it works but is kind of suboptimal
09:02:49 -!- immibis has quit (Read error: 60 (Operation timed out)).
09:02:56 <AnMaster> (it reads one char at a time then appends those to a string buffer)
09:02:59 <fizzie> G is used only at startup when it loads the state-file. It was easiest to do line-delimited stuff there.
09:03:10 <AnMaster> fizzie, I guess it works well enough for that
09:04:03 <AnMaster> ais523, and "ineffektivt" in Swedish means "inefficient"
09:04:35 <ais523> "I see what you mean now"
09:04:36 <AnMaster> like you knew it but had momentarily forgot it?
09:04:48 <ais523> so it was a case of remembering, too
09:04:58 * AnMaster don't remember having mentioned it
09:05:26 <ais523> you made the same mistake a few weeks ago, and I corrected you then, IIRC
09:06:00 <AnMaster> fizzie, anyway did you do a release build or a debug buil?
09:07:08 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | For outputting BugSophia: Gimme a RUNIC LETTER STAN!.
09:10:57 <fizzie> A release build, I think; I just bzr'd your thing, then ran cmake with -DUSE_64BIT=OFF (I don't really need 64-bit addressing and that box is an oldish Pentium M) and -DJAIL=ON (the chroot thing).
09:11:51 <AnMaster> actually that may be something in between
09:12:16 <AnMaster> since CMAKE_BUILD_TYPE wouldn't be set
09:12:51 <AnMaster> considered release by cfunge's "clean up on exit to please valgrind code"
09:13:46 <fizzie> If the cmake-generated flags.make has all the cflags, then it indeed doesn't seem to have -O2 in it.
09:14:09 <AnMaster> fizzie, cfunge builds and works fine at -O3 -fweb even here
09:14:56 <AnMaster> but I'm unsure about flags.make
09:15:13 -!- toBogE has quit (Read error: 110 (Connection timed out)).
09:15:26 <AnMaster> C_FLAGS = -pipe -march=k8 -O2 -msse3 -O3 -DNDEBUG -fweb -I/home/arvid/src/cfunge/trunk/src ....
09:15:38 <AnMaster> and C_FLAGS = -pipe -march=k8 -msse3 -ggdb3 -I/home/arvid/src/cfunge/trunk/src ...
09:18:29 <AnMaster> fizzie, so it could probably be even faster :D
09:19:05 <AnMaster> for that "game of life in befunge 93" -O3 makes a *LOT* of difference
09:19:11 <AnMaster> however it will be harder to catch bugs
09:19:50 <fizzie> Okay, I did -O3 now, just in case. (Actually I just ran ccmake on the build dir and changed the build type to Release, it seems to have added at least -O3 there.)
09:20:04 <ais523> why do people have difficulty debugging at O3, by the way?
09:20:09 <AnMaster> fizzie, you may want to set -march
09:20:22 <ais523> yes, it makes the code jump around a lot
09:20:37 <ais523> and you can't always variable queries directly
09:20:43 <ais523> but normally there's some way to get at the value
09:20:51 <AnMaster> ais523, x86 asm dump, register dump, sure whatever
09:20:52 <ais523> like evaluating an expression that's in the code
09:21:09 <ais523> e.g. if you're calling a function with the arg (y*2)+1
09:21:13 <AnMaster> (since I find x86 asm bloody awful)
09:21:14 <ais523> then p (y*2)+1 normally works
09:21:26 <fizzie> So, how do I add extra CFLAGS to that thing easily?
09:21:39 <AnMaster> fizzie, in ccmake you hit t for "advanced screen"
09:22:02 <AnMaster> fizzie, called CMAKE_CFLAGS or something like that iirc
09:22:58 <AnMaster> the resulting cflags will be CMAKE_C_FLAGS + CMAKE_C_FLAGS_${YOUR BUILD TYPE IN UPPERCASE}
09:23:06 <fizzie> Must restart that thing, then.
09:23:13 -!- fungot has quit ("flaggity flag").
09:23:23 <AnMaster> fizzie, well yeah cfunge doesn't do hot code swapping ;P
09:24:00 -!- fungot has joined.
09:24:13 <fizzie> At least the brainfuck infinite-loop was faster.
09:24:17 <fizzie> And that, too, I guess.
09:24:26 <AnMaster> didn't really dee any difference
09:24:48 <AnMaster> ah yes, a bit over 10 seconds before. less than 10 now
09:24:56 <fizzie> This time five seconds (as seen from my viewpoint; and we're on the same server with fungot). Last time it was 14 seconds.
09:24:57 <fungot> fizzie: i'll make one. when does sxyz ignore its argument?
09:25:06 <AnMaster> fizzie, well for the "game of life" for example, as I said it makes huge difference
09:25:40 <AnMaster> now that was quite a bit faster
09:25:55 <AnMaster> but it isn't in bf interpreted by befunge either
09:25:57 <ais523> AnMaster: thutubot's Underload interp is written in Thutu, rather than being written in BF written in Funge...
09:26:15 <ais523> ofc neither is particularly efficient, but I suspect thutubot's method is faster
09:26:37 <fizzie> The stand-alone Funge-98 Underload seems pretty fast.
09:26:44 <ais523> and thutubot got a lot faster once I started storing the timeout counter in balanced binary rather than in unary
09:26:47 <fizzie> Should just stick it into the bot.
09:27:08 <AnMaster> considering it may have fixed ideas about temp storage locations
09:27:24 <fizzie> Yes, a few numbers need to be changed.
09:27:38 <fizzie> Still, I wrote it with the intention of sticking it into the bot sooner or later.
09:27:39 <ais523> AnMaster: binary with digits 1,0,-1
09:27:47 <ais523> it uses 1 and 0 normally
09:27:54 <AnMaster> ais523, well so does computers
09:27:56 <ais523> but once the count reaches 0 it starts filling it with -1s
09:28:07 <ais523> because it's easier to figure out when it reaches the end that way
09:28:11 <ais523> I use the -1s for carry propagation
09:28:19 <ais523> well, borrow propagation
09:28:29 <ais523> because arithmetic via regex is hard
09:28:52 <ais523> Thutu arithmetic is probably easier than INTERCAL arithmetic, though, still...
09:29:21 <Deewiant> AnMaster, fizzie: just run inside a {, that's the main reason the instruction exists :-P
09:30:08 <oklopol> hmm, probabilities, and satisfaction don't really mix in my head
09:30:09 <ais523> hmm... a sandbox fingerprint?
09:30:15 <fizzie> Deewiant: I don't like {; it gets so messy to access not-related-to-the-code storage then.
09:30:17 <oklopol> *probabilities and satisfaction
09:30:20 <AnMaster> 1) make = in efunge evaluate erlang expressions 2) Make = mark the program as "tainted", somewhat like when you load a binary module into the linux kernel
09:30:39 <oklopol> i'm having a hard time even implementing just the brute-force way to interpret noprob
09:30:41 <Deewiant> fizzie: well, you're not supposed to do that :-P
09:30:43 <fizzie> Deewiant: Plus I don't like the fact that it sets storage offset to ip+delta; is there even a comfortable way of specifying an arbitrary storage offset?
09:30:48 <ais523> AnMaster: can you untaint the program by matching it against a regex?
09:30:52 <ais523> that works for Perl...
09:31:02 <Deewiant> fizzie: {, followed by u to push what offset you want, followed by }
09:31:03 <fizzie> Deewiant: There's the input given to the program, for one thing. And the output going out. I don't want to copy that around all the time.
09:31:10 <ais523> AnMaster: in Perl tainted data is data that comes from an unknown source
09:31:19 <ais523> the only way to untaint it is to run it through a regex
09:31:24 <AnMaster> ais523, ah well I meant in the same meaning as the linux kernel
09:31:27 <ais523> captured groups from the regex are untainted
09:31:29 <Deewiant> fizzie: I think you could push the output via u unless you're mutating it a lot
09:31:44 <Deewiant> and the input you could place inside the { area
09:31:50 <ais523> oklopol: because the Perl interp assumes you know what you're doing
09:31:52 <AnMaster> ais523, which is "binary/closed source module has been loaded, any crash backtrace will be unusable for kernel developers"
09:31:58 <ais523> and that the regex was defined to check that the input was safe
09:32:29 <ais523> so you can just do $unsafeinput =~ /^(.*)$/; $safeinput=$1;
09:32:37 <ais523> if you know anything but newlines are safe
09:32:46 <ais523> if you were expecting a number you do $unsafeinput
09:33:09 <ais523> *if you were expecting a number you do $unsafeinput =~ /^(-?[0-9]+)$/ or die; $safeinput=$1;
09:33:15 <ais523> that way you get an untainted number or an error
09:33:16 <fizzie> Deewiant: Okay, u-output could work. I just like the well-specified absolute addresses more. On the other hand, if this ATHR thing ever gets off the ground and I want to run interpreters concurrently, I guess I'll sort-of need to use the storage offset to help the different instances coexist.
09:33:17 <oklopol> and you can't do anything but matching with an unsafe string?
09:33:31 <ais523> you can do most things but the result is tainted
09:33:37 <ais523> e.g. concatentating it, or whatever
09:33:53 <ais523> there are various things you can't do: you can't pass it as input to a shell command, for instance
09:34:00 <fizzie> Deewiant: Actually using 'u' didn't occur to me; I'm more of a Befunge-93 person.
09:34:06 <AnMaster> ais523, what if you want to do it with something else than regex? Maybe write your own code to check it in a custom way, like is the normal way in for example C?
09:34:26 <ais523> AnMaster: well you just pass it to a regex at the end to tell Perl you've checked it
09:34:30 <ais523> it needs checking, and it needs a regex
09:34:36 <ais523> the checking doesn't have to be in the regex though
09:34:47 <AnMaster> ais523, that seems rather silly to me
09:35:04 <AnMaster> what about matching but keeping it tainted?
09:35:14 <ais523> you can taint data whenever you like
09:35:29 <AnMaster> ais523, anyway see above for what I actually meant
09:35:34 <ais523> yes, I know what you meant
09:35:42 <oklopol> the problem with variables with a probability in noprob is, you can't just store the probability, becauce vars may not be independent
09:36:03 <ais523> incidentally, doesn't y have a code for meaning "= has the same meaning as eval in the lang this interpreter is written in"?
09:36:32 <ais523> oklopol: ah, I know what you're getting at
09:36:40 <ais523> that's the same problem people have simulating a quantum computer
09:36:53 <ais523> you have lots of probabilities that depend on each other, so it's a pain to simualte
09:37:07 <ais523> except with a quantum computer it's worse because the probabilities are complex numbers
09:37:22 <fizzie> Regex is not the only way, though: "Values may be untainted by using them as keys in a hash; otherwise the only way to bypass the tainting mechanism is by referencing subpatterns from a regular expression match."
09:37:34 <ais523> ah, I forgot the keys in a hash one
09:37:42 <ais523> although I don't really get why it exists
09:37:43 <oklopol> ais523: but i was thinking, perhaps i should store everything the variable depends on and their independent probability, and check if vars are independent with an intersection
09:37:53 <AnMaster> did anyone understand xkcd today?
09:38:07 <oklopol> so i can use math in most cases, because usually the vars do *not* depend on each other
09:38:10 <ais523> oklopol: that sounds like an optimisation, does it solve the underlying problem though?
09:38:49 <fizzie> ais523: So you can untaint $x with the very pretty $x = (keys %{{ $x => 1 }})[0]; construct.
09:38:52 <oklopol> ais523: if the vars aren't independent, i will take all random vars the variables depend on, and enumerate all possible settings for them
09:39:06 <oklopol> and check what the probability is for each pair
09:39:07 <ais523> fizzie: does that even need the keys in there
09:39:27 <ais523> $x = (%{{$x=>1}})[0] would work, wouldn't it?
09:39:48 <fizzie> Yes, I think it would.
09:39:54 <ais523> or would it need to be written $x= @{[%{{$x=>1}}
09:40:00 <fizzie> The intention is clearer with the "keys". :p
09:40:00 <ais523> or would it need to be written $x= @{[%{{$x=>1}}]}[0]
09:40:07 <ais523> Perl casts are great...
09:40:25 <ais523> fizzie: that keys is inefficient!
09:43:43 <fizzie> AnMaster: "Burma-Shave was an American brand of brushless shaving cream, famous for its advertising gimmick of posting humorous rhyming poems on small, consecutive highway billboard signs."
09:44:24 <oklopol> what's it called when you use those funky trees to get optimal prefix-codes for tokens?
09:45:06 <fizzie> oklopol: Huffman coding?
09:46:25 <ais523> there's Shannon coding too but normally it doesn't work as well
09:46:36 <ais523> but in theory, you could end up with neither of them being optimal
09:47:04 <ais523> 90% 1 vs. 10% 0, the most optimal method is not to give them one bit each
09:47:18 <ais523> you could have, for instance, 00=0, 01=1, 1=111
09:47:29 <ais523> but I don't think Huffman coders take that sort of thing into accoutn
09:47:30 <oklopol> well that's a different coding scheme
09:47:44 <ais523> oklopol: well, isn't that proof that Huffman isn't always optimal?
09:47:47 <oklopol> no. of course you can get better compression given more sophisticated schemes
09:47:54 <oklopol> ais523: nothing is always optimal
09:48:17 <oklopol> huffman is optimal if you only have codes for *single tokens*, and codes are static
09:48:32 <oklopol> but, shannon is *never* optimal
09:48:37 <ais523> markov-chain codes tend to do better
09:48:46 <ais523> it's only optimal when it has the same result as Huffman
09:48:49 <oklopol> of course it's sometimes optimal, but it's never always optimal :P
09:50:37 <AnMaster> I have an optimal compression algorithm for some common data in esolang contexts
09:51:12 <AnMaster> most significant bit set = "Hello, World!", most significant bit not set = 99 Bottles of beer. Any other value is ignored.
09:51:38 <oklopol> "any other value is ignored"? :D
09:52:03 <oklopol> please show example of ignored encoding
09:52:04 <ais523> AnMaster: that's very HQ9+-like
09:52:04 <AnMaster> oklopol, well yeah, I didn't say it could encode everything did I?
09:52:22 <AnMaster> oklopol, as in "any other bit but the MSB is ignored"
09:52:23 <fizzie> oklopol: Most significant bit. I assume it's a multi-bit number we're talking about.
09:52:42 <ais523> hmm... is "Hello, World!" a quine in HQ9+...
09:52:44 <oklopol> i'm just being pedantic here, fizzie
09:52:54 <fizzie> oklopol: Yes, I'm just being slow. :p
09:52:59 <oklopol> 1 = hw, 0 = 99bob, X = ignored
09:55:36 <fizzie> 'X' - strong drive, unknown logic value
09:56:19 <oklopol> it shall not take any value for an answer!
09:57:07 <ais523> oklopol: high impedance is Z
09:57:17 <ais523> fizzie: actually, normally I treat "X" as meaning "short circuit"
09:57:26 <ais523> because it's what you get if you strongly drive something to both 0 and 1
09:57:40 <AnMaster> oklopol, I think they are thinking VHDL
09:57:46 <ais523> yes, 9-valued booleans
09:57:58 <oklopol> AnMaster: well i'm thinking just general circuits
09:58:24 <ais523> anyway, even if you aren't an electronic engineer I think you can appreciate that strongly driving the same part of the circuit simultaneously to 0 and 1 is a bad idea
09:58:58 <AnMaster> and what about driving faster than speed limit?
09:58:58 <ais523> weak driving is via a resistor
09:59:03 <ais523> whereas strong driving is direct
09:59:14 <ais523> in theory, strongly driving something to 0 is like connecting it to the negative power supply
09:59:26 <ais523> and strongly driving something to 1 is like connecting it to the positive power supply
09:59:43 <ais523> whereas weak driving can be overriden via strong driving
09:59:53 <ais523> strong 0 + strong 1 = short circuit (X)
09:59:59 <ais523> strong 0 + weak 1 = strong 0
10:00:13 <AnMaster> so what do you get when you connect something to both? Say a lamp to both ends of a battery?
10:00:13 <ais523> weak 0 + weak 1 = weak 0.5, or some other non-integral value (W)
10:00:25 <ais523> AnMaster: the lamp itself is a resistor there
10:00:41 <ais523> basically whatever you put between 0 and 1 either has to block the current, or use up the energy flowing through it
10:00:52 <ais523> if you put a wire there, it can't block the current because it's a wire and designed not to
10:00:59 <ais523> so it has to use up the energy flowing through it
10:01:05 <ais523> which it normally does by becoming very hot and melting
10:01:08 * AnMaster puts a diod the wrong way around between ais523's 1 and 0
10:01:24 <ais523> AnMaster: which way's "the wrong way"?
10:01:30 <ais523> reverse, so it blocks the current?
10:01:37 <ais523> or forward, so it doesn't block the current and catches fire?
10:01:46 <ais523> arguably reverse polarity is the safer one here
10:02:11 <fizzie> On the other hand it doesn't accomplish anything.
10:02:23 <AnMaster> hopefully that should give you radio Moscow (iirc)
10:02:24 <fizzie> At least the one catching fire is doing something.
10:02:30 <ais523> it does actually, because you can put an ammeter across the diode and determine how leaky it is
10:02:34 <AnMaster> ais523, and iirc you read the book so you should get that reference
10:02:36 <ais523> which is useful if you're into measuring diodes
10:02:46 <fizzie> Misread "because you can put an AnMaster across the diode and determine how leaky it is"
10:03:02 <ais523> measuring diodes is more useful than setting them on fire, IIRC
10:03:05 <fizzie> I'm not sure which one is the leaky one here.
10:03:28 <fizzie> ais523: If you *recall* correctly? So, uh, you read that in a book?
10:03:36 <AnMaster> ais523, didn't get the reference?...
10:03:38 <ais523> I meant IMO rather than IIRC, I think
10:03:57 <ais523> AnMaster: I almost got it, but wasn't sure what it was
10:04:05 <ais523> I thought "there's a reference to something there"
10:04:22 <ais523> still don't get the reference, although I know it exists
10:04:25 <AnMaster> ais523, if you still can't figure it out you need to re-read that book
10:04:38 <ais523> and yes, everyone needs to re-read Good Omens
10:04:40 <ais523> as many times as possible
10:04:49 <AnMaster> ais523, I have 5 times or so at least
10:04:54 <fizzie> I just read it not many weeks ago; do I _really_ have to reread it again?
10:05:02 <ais523> I've only read it about 3 times so far
10:05:09 <AnMaster> fizzie, you don't get the reference either!?
10:05:48 <fizzie> AnMaster: I assume it's something Newton Pulsifer did; but ais523 said everyone needs to reread it anyway.
10:07:08 <AnMaster> fizzie, the book mentioned some magazine once published a "joke circuit schema" that wouldn't work, but when Newton Pulsifer built it, it received Radio Moscow
10:07:16 <fizzie> Yes, I remember that bit.
10:07:21 <ais523> ah yes, I remember that now
10:07:27 <AnMaster> it also mentioned "diods the wrong way around"
10:08:15 <AnMaster> really you need to work a bit on your Pratchett trivia knowledge ;P
10:09:04 <AnMaster> also it may have been some other station than "Radio Moscow", not 100% sure there
10:10:08 <fizzie> "It had diodes the wrong way round, transistors upside down, and a flat battery. Newt had built it, and it picked up Radio Moscow."
10:13:39 <oklopol> fucking unimplementable :|
10:13:52 <ais523> oklopol: hmm... Feather makes me feel the same way, or worse
10:13:58 <ais523> because I'm almost convinced it is implementable
10:14:06 <ais523> I just find it really hard to figure out how
10:14:38 <oklopol> the problem is, usually i can at least implement some incredibly slow brute-force interpreter
10:14:51 <oklopol> but in this case that's simply excruciatingly hard
10:26:05 -!- oerjan has joined.
10:37:28 <fizzie> I don't know, it came from a .rtf file in that same pile I used for fungot's discworld training. :p
10:37:29 <fungot> fizzie: factloop is recursive, but it's handy.
10:37:43 <AnMaster> fizzie, Good Omens aren't discworld
10:37:59 <fizzie> Yes, I meant the same "around a thousand ebooks" pile.
10:38:25 <AnMaster> also a thousand... that's a lot of ebooks
10:38:33 <fizzie> None of them legal is my guess.
10:38:55 <fizzie> 600 megabytes; got that as a CD several years ago from a friend.
10:39:25 <AnMaster> fizzie, what dict is fungot on now?
10:39:26 <fungot> AnMaster: mwahah life is good. it is
10:39:44 <fizzie> And he certainly seems to enjoy it.
10:39:58 <fungot> ais523: if you email the mit scheme ref. google just turns up mailing list posts
10:40:01 <optbot> ais523: El dato que nada importa.
10:40:15 <ais523> hmm... thutubot doesn't react when I mention its name
10:40:36 * ais523 wonders how long it will be before optbot starts returning fungot-generated stuff
10:40:36 <optbot> ais523: of course +1 is wrong :)
10:40:36 <fungot> ais523: when i came across a program once control will never re-enter that point, then later hit i to go into grad school
10:40:58 <fizzie> Optimally you need a third paradigm for babble-generation; optbot and fungot are slightly different, after all.
10:40:59 <fungot> fizzie: yeah, that doesn't halt?
10:40:59 <optbot> fizzie: Of course as far as I know I'm the only person to use this pronoun.
10:41:19 <AnMaster> fizzie, yes, but what one would that be?
10:41:30 <ais523> markov-chaining the letters?
10:41:30 <fizzie> And I think it's only fair fungot's the one who's more messed-up in the head.
10:41:30 <fungot> fizzie: should be " prefix", and that functionality which can be adjust with forms like this:
10:41:47 <oerjan> hm fungot needs a corpus made from famous speeches
10:41:47 <fungot> oerjan: fnord firefox doesn't have bugs? will paredit be classified as an esolang
10:41:50 <ais523> we had alicebot vs. fungot a while ago IIRC
10:41:51 <fungot> ais523: you know if the source code
10:41:59 <ais523> but I was trying to think up something unusual
10:42:14 <oerjan> "Fnord and seven years ago ..."
10:42:17 <AnMaster> To be or not to fnord, kindom for a fnord
10:42:21 <ais523> an order-3 markov chain on letters not words would probably result in absolute nonsense, with no advantages other than being pronouncable
10:42:30 <fizzie> The lovecraft texts generated a _lot_ of fnords.
10:42:33 <AnMaster> oerjan, what would that be from?
10:43:10 <fizzie> fungot: Do some Cthulhu stuff again for a bit.
10:43:10 <fungot> fizzie: i tried not to heed him; tried to break through the paralysis which held me, and now and then to burst forth in a fnord anywhere that afforded me an opportunity to be near the college, and am fnord of get'g, there be'g ii. goode fnord in towne, dr, bowen and sam: fnord.
10:43:47 <AnMaster> and get'g well yeah both ' and g gets from funge space
10:43:55 <ais523> Thutubot wouldn't really be suited for random gibberish though
10:44:01 <ais523> as Thutu has no random number function
10:44:09 <oerjan> oh actually, "Four fnord and seven years ago ..."
10:44:13 <ais523> and no really sensible way to store data but the source code
10:44:16 <fizzie> "be'g" as in "being" in a folksy way.
10:44:51 <ais523> Thutu would probably be good at that
10:44:54 <ais523> but I don't like alicebots
10:45:03 <fizzie> I don't think I have a lot of speeches available right now. The Europarl speeches generated quite nonsensical stuff.
10:45:09 <ais523> they always sound stupid and aritifical
10:45:18 <AnMaster> ais523, I seen some pretty good ones
10:45:30 <ais523> and they just keep on talking on the same subject as me
10:45:35 <ais523> people never do that in real life
10:45:41 <AnMaster> that can keep track of different topics for different speakers
10:45:56 <AnMaster> ais523, so you need to mess around with that a bit so it sometimes changes topic
10:45:58 <ais523> also, the vocabulary tends to be limited to mine
10:46:19 <ais523> probably why I don't like alicebots is obvious if you just pipe two of them into each other
10:46:27 <ais523> the result tends to be obviously awful
10:46:40 <ais523> because both of them are trying to generate a conversation off what the other says
10:46:44 <ais523> whilst being non-commital themself
10:46:53 <ais523> because they don't know about any subjects themself...
10:46:54 <oerjan> AnMaster: http://en.wikipedia.org/wiki/Gettysburg_Address#Lincoln.27s_Gettysburg_Address
10:47:14 <ais523> oerjan: Wikiquote will be better for Wikipedia for that sort of thing, probably
10:47:21 <AnMaster> ais523, what about mixing that up with some markov stuff?
10:47:36 <fizzie> Okay, I fed it everything in my gutenberg pile having the word "speeches" in the title. :p
10:47:45 <fizzie> fungot: Be pompous for us, please.
10:47:45 <fungot> fizzie: in the mean time, it was done. before that, two indians were placed on the council of the fnord fnord
10:48:08 <fizzie> Ooh, the council of the fnord fnord.
10:48:12 <ais523> that makes sense apart from the fnords
10:48:56 <fungot> AnMaster: i consider this, sir, i greatly deceive myself, that the judge is to hear the government accused of avoiding the discussion of the right honourable baronet proposes to punish brazil for the slave trade, not in themselves presumptively criminal, but actions neutral and indifferent the whole matter, in which mode of government, there is an archbishop with ten thousand a year, he has done for him more than the ordinary b
10:48:56 <ais523> actually, I've had a brilliant idea for something to Markov-chain
10:49:06 <AnMaster> fizzie, you just changed to EU one?
10:49:31 <ais523> if you markov on compressed text, presumably it'll still have the right format, apart from maybe checksums
10:49:31 <fungot> AnMaster: yours very sincerely and respectfully, abraham lincoln, fnord
10:49:49 <fizzie> Here's the list: http://zem.fi/~fis/speeches.txt
10:49:55 <ais523> and decompressing it will presumably lead to something with a coherent subject
10:50:03 <oerjan> oh so lincoln was a fnord too
10:50:03 <ais523> because the same things will be said on many occasions
10:50:11 <fungot> AnMaster: " there are who, while to vulgar eyes they seem of all my honours, i am sure that my colleagues will not fnord, seeing that there is a fnord, is almost totally wanting, and then you will be prepared against these inconveniences.
10:50:52 <AnMaster> ais523, how do you know where to cut it off then?
10:51:05 <ais523> AnMaster: include EOF in the markov chain, obviously
10:51:28 <AnMaster> ais523, no I mean would you be able to cut off after any byte?
10:51:35 <AnMaster> I don't know the format for gzip well
10:51:39 <ais523> yes, you markov-chain the bytes
10:51:56 <AnMaster> ais523, would that really work?
10:52:02 <ais523> at the highest order you can get away with without verbatimising
10:52:08 <ais523> and almost certainly not, but it would be great to try
10:53:22 <ais523> going for a while, I need to get food
10:53:30 <fizzie> Misread "I need to get fnord".
10:53:37 <fizzie> Thanks to all the fnords here lately.
10:53:53 <fizzie> Haven't tried Shakespeare, no. He's not in the preprocessed Gutenberg books I have for that other project, so it'd need some text-extraction.
10:55:13 <AnMaster> fizzie, http://en.wikisource.org/wiki/Hamlet
10:55:15 <fizzie> Not sure where I could get conversational text easily, except the IRC logs. (And the telephone conversations, but I tried that already.)
10:55:55 <AnMaster> fizzie, you could filter it to just be the actual phrases said, removing all "enter Hamlet" "exit hamlet"
10:56:02 <fizzie> Actually wait, Shakespeare is there.
10:56:47 <fizzie> http://zem.fi/~fis/ss.txt
10:57:01 <fizzie> I guess I could try that too, although it probably won't be very interesting.
10:57:24 <oerjan> the fnordity should be staggering
10:57:34 <AnMaster> fizzie, would be nice to try at least
10:59:13 <fungot> fizzie: ' so in thyself thyself art made away; a mischief worse than civil home-bred strife, fnord or, hubert, hubert throw thine eye on yon young boy: ile tell thee more.
10:59:30 <AnMaster> fizzie, how does it handle ' ?
10:59:37 <fungot> AnMaster: ped. keepe your hundred pounds to yourself: he shall draw; he shall tap; said i well, bully hector?
10:59:50 <fungot> AnMaster: iohn. i will presentlie goe learne their day of marriage, and things, and to confine yourself to fnord house convey our arms, quartered within his fnord th' expense of many a blasting hour, let it no more on height of our care, sir.
10:59:59 <fizzie> Inside words, ' is just part of the token; so "foo's" is one token while "foo" is another.
11:00:29 <fizzie> Things that look like "'foo bar baz'" get turned into "opening single quote, foo, bar, baz, closing single quote".
11:00:39 <AnMaster> fizzie, You are unusually random aren't you?
11:00:48 <AnMaster> fungot, You are unusually random aren't you?
11:00:48 <fungot> AnMaster: brutus. enough, i warrant you
11:01:02 <fungot> AnMaster: louel seemes to stay. you, in the meane time, some show in the posterior of the day
11:01:14 <fizzie> I'm not sure what it does with "'tis" and "th'". Probably interprets "'tis .. th'" as a single-quoted string.
11:01:51 <fungot> optbot: laf. and shall i flye? i haue told thee, of a most homely fnord man, aufidius, piercing our fnord then we thought them none, her eie is sicke on't, i would
11:01:52 <fungot> optbot: ant. now my spirit is thine the better part made mercie, i should
11:01:52 <optbot> fungot: we could add type inference
11:01:52 <fungot> optbot: as through an arch so hurried the blown tide as the recomforted through the gates of millaine, and ith' dead of darkenesse the ministers for th' present. get thee to bed.
11:01:52 <optbot> fungot: that's where I stop agreeing :P
11:01:53 <fungot> optbot: iago. i'll send him ( for so i thought i had, i neuer spent an houres talke withall. his eye is hollow, and hee must needes goe that the diuell driues
11:01:53 <optbot> fungot: you were all hardcore anti scheme
11:02:10 <oerjan> i think optbot didn't like that
11:02:11 <optbot> oerjan: Glyph means symbol, right? (or atleast something similar)
11:02:32 <fizzie> It looks shakespeareanistic, but all in all not terribly funny.
11:03:44 <fizzie> Heh, 514 megabytes of language models already. (Lewis Carroll, Darwin, Europarl, telephone calls, IRC logs, lovecraft, those speeches and now Shakespeare.)
11:03:45 <AnMaster> fizzie, well this will be fun when it gets into optbot
11:03:46 <optbot> AnMaster: so no need to pack more closely
11:04:18 <AnMaster> fizzie, I got to say irc is the one that worked best so far
11:04:36 <AnMaster> fizzie, what about that idea with wikipedia talk pages?
11:05:04 <AnMaster> fizzie, would be nice to see how ais reacts when he get back on that too :D
11:05:12 <fizzie> Maybe. Should probably check out whether the pre-supplied Wikipedia data dumps include talk page contents.
11:05:28 <fungot> AnMaster: then there is schem48...
11:05:30 <optbot> AnMaster: With practice, dream recall can be "learned".
11:05:30 <optbot> fungot: cygwin is yer friend
11:05:30 <fungot> optbot: thinking huh? maybe file it upstream?
11:05:31 <optbot> fungot: in that case, you've missed the joke entirely
11:05:31 <fungot> optbot: i've never seen that problem before, which is odd
11:05:31 <optbot> fungot: Of course predictably it still doesn't work :|
11:05:32 <fungot> optbot: sometimes watching the politics here, it was used
11:05:32 <optbot> fungot: !pager %a A C T I O N s h o w s S i m o n R C%a
11:05:32 <fungot> optbot: fnord/ fnord it was simply suggested that there should be
11:05:42 <fizzie> Cleaned up the files a bit.
11:05:54 <fizzie> "pages-full.xml.bz2/7z - Current revisions, all pages (includes talk and user pages)"
11:06:06 <fizzie> For some reason they don't do database dumps including only talk pages. :p
11:07:10 <fizzie> Trying to figure out. Apparently the dumps aren't exactly those as the wiki page describes.
11:07:34 <fizzie> "Discussion and user pages."
11:08:03 <fizzie> Yes. Talk pages should contain at least a bit of chatter.
11:09:20 <fizzie> I'd like to fetch the talk pages only, but apparently there's only "articles without talk" and "articles, user pages and talk" dumps.
11:10:40 <AnMaster> could take quite a while to download
11:10:46 <fizzie> I wonder if that's worth trying.
11:11:00 <AnMaster> fizzie, also you need to restrict yourself to a subset of the talk pages
11:12:00 <fizzie> ETA 2 hours, apparently.
11:12:54 <fizzie> Around 1.7 megabytes / sec. Not too shabby, although far from the speeds I used to get back when I still lived on the university student housing place.
11:14:32 <fizzie> Maybe I should look around our department's file system; so far I've just looked at the speech group's "text" directory.
11:15:32 <fizzie> At least the natlang people have a "corpora" directory.
11:16:13 -!- oklopol has quit (Read error: 113 (No route to host)).
11:19:45 <fizzie> It's that XML MediaWiki export format... I guess I can pipe the file through bunzip2 and some script to output only the interesting things, so I don't need to actually uncompress that 8-gigabyte file.
11:21:42 <fizzie> No, I'll probably need to tweak the script so much that it's better to have a local copy.
11:21:49 <fizzie> But bzcat's pretty much "bunzip2 -c".
11:22:05 <fizzie> Er, I mean, "bzip2 -dc".
11:22:58 <AnMaster> fizzie, how large would the expanded file be?
11:23:02 <fizzie> I'm sure I'll find some use for a local Wikipedia dump other than just fungot, sooner or later. (Although it'll be pretty dated fast.)
11:23:02 <fungot> fizzie: i also put an fnord man as the child of a, having b fnord a a fnord maybe b
11:23:57 <AnMaster> fizzie, if you update the irc model remember to filter fungot itself first please
11:23:58 <fizzie> Do the bzip2 headers include uncompressed size? The wiki-dump-download-site doesn't say.
11:23:58 <fungot> AnMaster: i camp at the bar and without the repeated fnord?
11:24:25 <AnMaster> anyway if you don't filter itself it would saturate I suspect
11:24:35 <AnMaster> a bit more every time it is updated
11:25:17 <AnMaster> [IpID || {_Key, IpID} <- Entries]
11:25:53 <fizzie> Apparently the uncompressed size is not stored in the bzip2 format (unlike gzip), so can't say. I would assume it's at the very least twice that size, probably a lot more.
12:02:42 <ais523> actually, if you want to test things on the Wikipedia dumps, one well-known trick is to get Simple English not English
12:02:50 <ais523> because you can still read it, and it's smaller
12:04:45 <fizzie> 56 % already fetched out of that enwiki dump, I guess I'll just select a couple of talk pages out of it or something.
12:16:28 -!- LinuS has joined.
12:21:03 <AnMaster> Deewiant, You wondered why R11B wouldn't work, well I read the changelog, and to be able to follow funge spec I need to use a function added in R12B-3
12:24:42 <Slereah_> If you put a multiple of the Chaitin constant in a check.
12:24:57 <ais523> not cash it, almost certainly
12:25:23 <Slereah_> Roughly probability of the stopping of a machine
12:25:29 <ais523> AnMaster: the probability that a random Turing machine halts
12:25:40 <ais523> it's been mathematically proved to be impossible to calculate accurately
12:25:48 * oerjan thought check formatting was fairly rigid
12:25:51 <Slereah_> Well, it's between zero and one.
12:25:53 <ais523> I'm not sure how many decimal places of it are known
12:26:04 <ais523> because it is possible to get bounds on it
12:26:21 <AnMaster> <ais523> it's been mathematically proved to be impossible to calculate accurately <-- transcendental(sp?) number?
12:26:31 <ais523> it depends on how you randomize the Turing machine, but I thought there was some official method
12:26:39 <ais523> AnMaster: trancendental numbers can be calculated sometimes
12:26:49 <ais523> pi is trancendental, but can be calculated to any number of decimal places
12:27:01 <AnMaster> well you said it is known to some decimal places?
12:27:05 <ais523> whereas for Chaitin's number there's some number of decimal places past which it can't be calculated, even in theory
12:27:19 <Slereah_> You can get an estimation, AnMaster
12:27:37 <Slereah_> It's a sum of 2^p * 0 if it doesn't stop, 1 if it does
12:28:08 <Slereah_> It's a probability, so you know it will be between 0 and 1.
12:28:16 <oerjan> basically it takes a while until you get to the first Turing machine you cannot decide, and before that you have got a few dozen bits iirc
12:28:23 <Slereah_> If the first program stops, you know it's between 0.5 and 0
12:28:40 <AnMaster> 2^p * 0 well that would be (2^p) * 0, which would be 0?
12:28:52 <oerjan> oh and yeah shorter Turing machines are weighted more
12:28:57 <Slereah_> But if you can't prove anything about the first machine, you will always have an uncertainty of 0.5
12:32:20 <Slereah_> Maybe I should give out checks for floor[n*chaitin]
12:34:11 <ais523> I actually read part of Chaitin's autobiography; it seems he's a big fan of Lisp
12:34:23 <ais523> so much so that he once implemented a Diophantine equation that interpreted it
12:34:55 <ais523> restricted to nonnegative integers for each unknown
12:35:15 <ais523> so 3 / x = 2 has no solution as a Diophantine equation
12:35:26 <ais523> and "unknown" as a noun means one of the variables in an equation that you have to find the value for
12:35:34 <ais523> as opposed to a parameter, where you're given the answer in advance
12:35:53 <ais523> so if I say "solve x + 1 = y for y given that x is 3", then y's an unknown and x is a parameter
12:36:03 <AnMaster> hm... 2x=3 ..... x = 3/2 but yeah requires more than integers
12:36:21 <ais523> in general, solving Diophantine equations is super-TC
12:36:40 <ais523> because the equation itself can be TC, and a solution therefore solves the halting problem
12:37:05 <ais523> the Lisp one can easily be tweaked to implement a Turing machine by feeding it a Turing machine simulator written in Lisp
12:37:06 <AnMaster> ais523, ok. How do you make an equation TC though?
12:37:38 <ais523> AnMaster: if you give it certain input as parameters, the values the unknowns can take are a function of them which requires TC-ness to calculate
12:40:01 <oerjan> AnMaster: http://en.wikipedia.org/wiki/Hilbert%27s_tenth_problem
12:43:30 <oerjan> from the history section it's pretty clear that it is not a simple result
12:52:35 <oerjan> "Likewise, despite much interest, the problem for equations over the rationals remains open."
12:52:52 <oerjan> so no one knows if equations with fractions are TC
12:54:19 <AnMaster> ais523, do you know if pthreads threads share their working directory?
12:59:47 <AnMaster> It is implementation-defined if each thread got it's own working directory or if
12:59:47 <AnMaster> the working directory is global for the whole implementation.
13:00:20 <fizzie> At least the POSIX specs speaks only of the "current working directory of the process".
13:00:26 <fizzie> Which would imply it's not thread-specific.
13:00:32 <AnMaster> fizzie, well in erlang it global per node
13:00:48 <AnMaster> so if you run the stuff distributed *some* threads may end up sharing directory
13:01:40 <AnMaster> assuming the nodes aren't set to use a common file server process on a single node
13:04:02 -!- Corun has joined.
13:07:27 <fungot> echo reverb rev bf rot13 hi rev2 fib wc ul ctcp
13:08:26 <fungot> oerjan: i've never been
13:08:27 <ais523> AnMaster: no I don't know about pthreads and directories, unfortunately
13:08:31 <fungot> echo reverb rev bf rot13 hi rev2 fib wc ul ctcp
13:08:57 <ais523> my guess is hi's defined to the 0-length string
13:09:09 <ais523> bf is only in that list because I defined a command called bf to try to confuse it
13:10:58 <oerjan> ^def pal bf >,[.>,]<[.<]
13:11:57 <oerjan> ^def pal bf >,[.>,]<<[.<]
13:14:05 <fizzie> Actually "hi" is an empty program.
13:14:10 <fizzie> That's why it won't show up.
13:14:26 <fizzie> Oh, you said that already.
13:14:43 <fizzie> I don't think it was actually empty, just not brainfuck so the bytecode compiler mostly ignored it.
13:15:25 <fizzie> That 'wc' program doesn't look like it should work.
13:15:30 <oerjan> ^def hi bf ,[.,]!Hello, World!
13:15:53 <fizzie> It doesn't accept preprepared input, I think.
13:15:57 <ais523> my guess is you can't specify input in defined commands
13:16:05 <ais523> hmm... maybe you should add Easy as a lang to fungot
13:16:06 <fungot> ais523: that result is reversed at the end where krishnamurthi got into a brief fnord with a normal distribtion, 95% are within 2 feet of it.
13:16:12 <ais523> despite being a joke originally, it's actually quite interesting
13:16:23 <ais523> but the input and program are in the same stream
13:16:33 <fizzie> It used to work, though. Defining 'hi' like that would've caused "^hi test" to output "Hello, World!!test".
13:16:46 <ais523> so for instance you can do ,H.,e.,l.,l.,o., .,W.,o.,r.,l.,d.
13:17:36 <oerjan> oh if the final input was added with ! ...
13:17:50 <fizzie> oerjan: Yes, I think. Because it would've executed "^hi test" by turning it into the original bf command with "!test" after it.
13:20:16 * ais523 wonders why eir inbox is now called INBOX rather than Inbox since eir email problems
13:21:15 <oerjan> the email client is hard of hearing so they had to shout the name to make sure it'll pick it up
13:21:24 -!- KingOfKarlsruhe has joined.
13:28:12 <ais523> wow, my latest FRC entry hit a bug in Google
13:29:11 * oerjan wonders if there was ever an FRC spam round, and if so if all messages got through
13:30:36 <ais523> well, it's probably the first FRC round with a word that's 200 kilobytes long
13:30:40 <ais523> made entirely out of Z in this case
13:30:56 <ais523> because the full chemical name for titin was mentioned a couple of entries earlier
13:31:03 <ais523> and each entry has to set a new record for word length...
13:34:57 <oerjan> so this may be the first round that is won due to technical issues? :D
13:35:17 <oerjan> (assuming it hasn't happened already)
13:35:26 <ais523> Google relayed the message fine
13:35:31 <ais523> its web-based view also displays it fine
13:35:35 <ais523> but has a "read more" link
13:35:40 <ais523> after the end of the message
13:35:52 <ais523> clicking on it gets you a "download full message" link, also after the end of the message
13:35:56 <ais523> and in both cases, the full message is shown
13:36:06 <ais523> it seems Google can handle 200KB messages, but not 200KB words
13:54:35 -!- oklocod has joined.
13:59:01 <AnMaster> ais523, odd, you only use e* for yourself instead of his in /me when it is about mail and nomic it seems
13:59:26 <ehird> what's wrong with gender-neutral pronouns
14:00:14 <AnMaster> ais523, what does that "FRC" you referred to mean? Googling for "define:FRC" lists lots of various things
14:00:16 <oklocod> a diophantine equation that runs lisp :D
14:00:20 <ais523> Agora uses "e" everywhere, when we're thinking about nomic we tend to slip into it
14:00:28 <ais523> AnMaster: Fantasy Rules Committee
14:00:39 <AnMaster> which was not one of the things google listed
14:01:44 <AnMaster> ais523, also 200 kb of Z really a valid word?
14:01:53 <ais523> they've allowed invalid words before
14:01:58 <ais523> just they have to make sense in context
14:02:10 <ais523> "Zzzz" is a common representation for sleep in English
14:02:15 <ais523> with variable numbers of Zs
14:02:22 <ais523> so I used 200,000 of them for Sleeping Beauty
14:02:41 <AnMaster> ais523, would it be invalid for someone else to use 300,000 such then?
14:03:03 <ais523> yes, but only because my rule made it invalid
14:03:42 <AnMaster> ais523, what is the full name of titin?
14:03:54 <ais523> AnMaster: about 180,000 characters long
14:04:21 <AnMaster> the link at wikipedia to a page with the full name wasn't valid
14:04:49 <ais523> Methionylthreonylthreonyl...isoleucine is how it's abbreviated on the wiki page about the longest word in English
14:05:16 <fizzie> They should abbreviate it as M189817E
14:05:30 <fizzie> In the I18N and L10N style.
14:16:20 <Deewiant> gawd, I hate those abbreviations
14:19:54 <AnMaster> abbreviation maybe? haven't checked length
14:19:57 <Deewiant> argumentation? approximation? assemblywomen?
14:20:18 <fizzie> oerjan: Antiquitarian faster-than-light?
14:20:53 <AnMaster> ehird, you didn't get the point of the joke...
14:21:07 <AnMaster> and it wasn't even my joke so hah
14:21:09 <ehird> i missed the part where it was funny.
14:21:21 <fizzie> egrep '^l.{10}n$' /usr/share/dict/words is a wonderful way of finding new meanings for it.
14:21:23 <ehird> (probably due to its nonexistance)
14:21:46 <Deewiant> fizzie: except here, where /usr/share/dict/words is swedish :-P
14:22:05 <ais523> here it's configurable
14:22:09 <ais523> so I set it to UK English
14:22:20 <AnMaster> wouldn't it be Finnish over there Deewiant?
14:22:59 <Deewiant> AnMaster: that's what I would have expected, that or English, but no, it's Swedish.
14:23:13 <fizzie> Here it's a symlink to /etc/dictionaries-common/words, which is a symlink to /usr/share/dict/web2.
14:23:36 <Deewiant> Well yeah, it's /usr/share/dict/words -> /etc/dictionaries-common/words -> /usr/share/dict/swedish
14:24:31 <AnMaster> fizzie, here it is just a file
14:24:32 <Deewiant> funny, american-english, british-english, and swedish, but no finnish
14:24:43 <oerjan> darn i counted wrong :D
14:25:26 <AnMaster> ais523, for that FRC, are you using /usr/share/dict/words normally?
14:25:34 * AnMaster wonders if there is any way to sort by length
14:25:48 <fizzie> I have /usr/share/dict/finnish from the wfinnish package, but it's a bit useless.
14:26:51 <fizzie> There are so many suffixes that it doesn't make much sense to try listing them all in a static word list.
14:27:04 <oerjan> <ehird> i missed the part where it was funny.
14:27:34 <oerjan> you and AnMaster are like a comical duo...
14:27:46 <ehird> i'm the funny one and he's the annoying one?
14:29:12 <AnMaster> oerjan, anyway it was Deewiant that made the joke
14:30:07 <Deewiant> it was more an explanation of my annoyance towards the abbreviations than an attempt at mirth
14:30:26 <AnMaster> hm gnu sort got an option called --random-sort, doesn't seem to be bogosort though
14:30:30 <AnMaster> -R, --random-sort sort by random hash of keys
14:30:30 <AnMaster> --random-source=FILE get random bytes from FILE (default /dev/urandom)
14:30:41 <ais523> I don't get what -R does at all
14:30:46 <ais523> hmm... maybe it sorts same lines together
14:30:47 <ehird> it shuffles the file
14:30:52 <ais523> but not in any particular order apart from that?
14:31:07 <ehird> it just shuffles the file.
14:31:32 <AnMaster> identical lines are put together though
14:31:52 <ais523> so it's taking hashes using a random but consistent algorithm, and sorting those
14:32:37 <oerjan> "hash of keys" so you don't necessarily use the whole line i guess
14:33:25 <AnMaster> -u, -r, -n, -i and -k are the only gnu sort options I use with any sort of regularity.
14:33:40 <AnMaster> oerjan, and yes, -k sets that iis
14:33:49 <AnMaster> -k, --key=POS1[,POS2] start a key at POS1, end it at POS2 (origin 1)
14:34:03 <AnMaster> a pitty you can't sort on field 3, then field 2 or such
14:34:21 <Deewiant> yeah, hence I usually write short haskell programs to do sorting :-P
14:34:30 <Deewiant> I suppose learning awk would be more optimal
14:34:48 <AnMaster> Deewiant, usually I use awk yes
14:35:08 <AnMaster> well I don't know perl myself, but I heard awk and perl were pretty similar
14:35:16 <ais523> Perl is a lot more featureful
14:35:20 <ais523> and awk compiles trivially to it
14:35:31 <AnMaster> ais523, awk is easier to write though IMO
14:35:40 <ais523> Perl is easy to write once you get the swing of it
14:35:42 <fizzie> The hashing is actually MD5 in GNU sort, and the randomness comes from the fact that it starts by hashing some random data.
14:36:27 <fizzie> Well, there's that random-source option.
14:36:38 <fizzie> It reads a bit from there and MD5-processes it first.
14:37:11 <fizzie> Then it takes the MD5 hashes and sorts according to those.
14:37:15 <AnMaster> also does it use the same random value for all the lines?
14:37:18 <Deewiant> so, given, say, data with 8 whitespace-separated columns per line, how would you sort it first according to the fifth and then the second column
14:38:01 <fizzie> If I'm reading it right, no. But it uses the same random value for both keys in all pairwise comparisons, so completely identical keys will be kept together.
14:39:35 <ais523> you can do it by running sort twice
14:39:36 <AnMaster> in awk: I would probably sort it using arrays in awk
14:39:46 <ais523> sort by the second, then sort the result by the fifth
14:39:49 <ais523> because GNU sort is stable
14:39:50 <AnMaster> ais523, sort(1) may not be stable iirc.
14:40:00 <ais523> ah, I thought it was a GNU sort question, which is IIRC
14:40:03 <AnMaster> ais523, only if given --stable
14:40:15 <AnMaster> otherwise why would they have that option?
14:40:17 <AnMaster> -s, --stable stabilize sort by disabling last-resort comparison
14:40:32 <Deewiant> what if the columns don't start at the same position, doesn't sort(1) require the exact column?
14:40:35 <ais523> ah, last-resort comparison isn't always used, IIRC
14:40:41 <ais523> and it would be stable if it wasn't
14:40:45 <AnMaster> Deewiant, field delimiters usually
14:40:47 <ais523> when it is you have the option to turn it off
14:40:53 <Deewiant> so, you might get "1 2 3 4 5 6 7 8" but also "1234 2 3 4 5 6 7 8"
14:41:09 <AnMaster> Deewiant, then if the field delimiter is space that should work
14:41:15 <Deewiant> okay, what if you're given "1|2,3,4|5 6 7 8" :-P
14:41:31 <AnMaster> Deewiant, then you give the option to set | or whatever you want as field delimiter
14:41:48 <AnMaster> which for gnu sort seems to be -t
14:41:49 <Deewiant> AnMaster: the field delimiter there is | twice, , twice, and space thrice.
14:42:06 <AnMaster> Deewiant, then you pipe it through sed first
14:42:23 <Deewiant> AnMaster: and what if | or , can appear in quoted strings
14:42:33 <ais523> then maybe sort is the wrong utility to be using?
14:42:34 <AnMaster> use some other char that won't
14:42:42 <Deewiant> AnMaster: and what if I can't affect the output data
14:42:54 <Deewiant> AnMaster: maybe @ can appear as well
14:43:02 <Deewiant> exactly, maybe sort is the wrong tool
14:43:05 <AnMaster> Deewiant, the reason for "maybe"
14:43:18 <AnMaster> you can select your own one there
14:43:32 <AnMaster> and of course there are cases when sort is the wrong tool
14:43:44 <AnMaster> no tool is optimal for every task
14:43:52 <ais523> Deewiant: if you're using that sort of format, it's likely some sort of CSV variant
14:43:54 <Deewiant> AnMaster: that isn't doable with sed, at least not easily.
14:44:03 <ais523> in which case OpenOffice would manage it just fine, except it isn't a command-line tool
14:44:14 <Deewiant> openoffice is a bit too heavyweight :-P
14:45:07 <AnMaster> Deewiant, why not? say you selected ! as separator, you can replace it with something else of course.... Then you do s/!/|/;s/!/,/;s/!/,/
14:45:17 <AnMaster> as long as you don't use g only the first one will be replaced
14:45:31 <Deewiant> AnMaster: say I'm given "1|2,3,4|5 6 7 8" where each number can contain quoted strings containing any of " |,".
14:45:51 <Deewiant> AnMaster: how do you sed the separators to something uniform so that you can turn it back later, without affecting the fields?
14:45:52 <AnMaster> Deewiant, well then you need a full scale parser
14:46:07 <AnMaster> which is possible in sed since it is TC, but I wouldn't recommend it
14:46:10 <Deewiant> or not really, that can be handled in regex.
14:46:21 <AnMaster> <AnMaster> you can select your own one there
14:46:21 <AnMaster> <AnMaster> that doesn't collide
14:46:21 <AnMaster> <AnMaster> and of course there are cases when sort is the wrong tool
14:46:21 <AnMaster> <AnMaster> no tool is optimal for every task
14:46:21 <AnMaster> <AnMaster> nor any programming language
14:46:38 <ais523> Deewiant: what if it contains email addresses, which aren't escaped, but it ignores separators inside comments in the email addresses due to knowing how to parse them?
14:47:13 <fizzie> That would be a really sensible format, ye.
14:47:17 <Deewiant> AnMaster: as I said above, what if the output format is not under my control
14:47:23 <AnMaster> what if it contains a hypothetical imaginary ASCII char with the value of sqrt(-1)?
14:47:33 <Deewiant> AnMaster: and my whole point was to ask what method/tool people would use
14:48:26 <fizzie> If someone gave me a file like that right now, I'd probably use a bit of Python, since it's got the 'csv' module (which does quoted strings just fine) in the standard distribution. I'm sure there's a CPAN Perl module, but maybe not installed by default.
14:48:31 <AnMaster> Deewiant, also sometimes you can sed back as I suggested above to get the right format
14:48:35 <AnMaster> and sometimes you want another tool
14:48:38 <fizzie> Of course that wouldn't help with the unescaped email addresses.
14:48:48 <AnMaster> Deewiant, like yacc + a C program
14:53:36 <AnMaster> Deewiant, anyway not sure if you saw it above, but by reading erlang news file I found that I need at least R12B-3 for efunge
14:53:54 <AnMaster> and there may be some bugs that affect in that version
14:54:42 <Deewiant> yes, I do notice when I am pinged
14:55:17 <ais523> so do I, but I'm pinged so often, it seems, that I'm beginning to grow to ignore them
14:55:53 <fizzie> I don't use it any more. :p
14:56:28 <fizzie> Although as a reference, I've got the comment "ping? pong!" there in the fungot sources where it answers a PING message.
14:56:28 <fungot> fizzie: perhaps the processor fried?
14:56:43 <fizzie> fungot: Huh? You seem to be working just fine.
14:56:43 <fungot> fizzie: i'd say another way to set some value to some specific variables and then using a foreign-lambda ( without) declaration instead.
14:56:56 <oerjan> fungot: can i have a burger with that?
14:56:56 <fungot> oerjan: in mornfall's future, that are just ' go to hell
14:57:12 <ehird> oh look, AnMaster made a joke by interpreting something as erlang
14:58:16 <AnMaster> Deewiant, does that mean anything?
14:58:29 <ehird> oerjan: opt-pong and opt-alt-pon
14:58:37 <AnMaster> oerjan, I'm not sure is päng ~ peng?
14:58:46 <AnMaster> maybe different spelling just?
14:59:04 <oerjan> oh so it's e in swedish too?
14:59:04 <ehird> ʼþ ʼß œ¨ʼþ´ ¯ ©øøð ´˜¸®¥,þʼø˜ ˛´¸Ë¯˜ʼ߲≤ ʼƒ ¥ø¨ Ëߨþ ˙¯˜þ þø ¸ø˜ƒ¨ß´ ,Ç¿,-´≥
14:59:07 <AnMaster> well not "coins", but rather the base of the word
14:59:41 <oerjan> Pengar har jag inga, men en sak til tröst
14:59:57 <AnMaster> for some reason "en peng" sounds strange...
15:00:11 <oerjan> it's "en penge" in norwegian
15:00:29 <oerjan> AnMaster: closer than expected then
15:00:31 <ehird>  ̑ø≤ ¸¯˜ ¯˜¥ø˜´ ®´¯ð þËʼß¿ ʼæ˛ ©¨´ßßʼ˜© ˜øþ≥
15:00:44 <ais523> ehird: stop pasting mojibake into the channel
15:00:59 <ehird> it's hold-down-opt-encoding.
15:01:07 <ehird> you hold down opt/alt in os x and type.
15:01:09 <AnMaster> oerjan, also it sounded like part of a poem? Since the grammar seemed strange
15:01:10 <ais523> Deewiant: Japanese written in UTF-8, but read as Latin-1
15:01:19 <ehird> ̛þæß œ¨ʼþ´ ´ƒƒ´¸þʼˇ´ ƒø® ¸ø˜ƒ¨ßʼ˜© ,ǿ,-´≥
15:01:21 <Deewiant> yes, I know, and I said it doesn't look like it :-P
15:01:23 <ehird> heh, some words are readable
15:01:24 <oerjan> AnMaster: Evert Taube's Flickan i Havanna
15:01:28 <ais523> and you're right, it doesn't look all that much like mojibake, it's hitting the wrong extended characters
15:01:45 <AnMaster> oerjan, Well he *was* Swedish, so that explains a bit I guess ;P
15:02:19 <AnMaster> oerjan, can't stand Taube really
15:02:29 <AnMaster> the music isn't good at all IMO, and the texts are worse
15:02:30 <ais523> that's why I was confused
15:02:37 <ais523> I was wondering if it was reverse mojibake or something
15:02:51 <AnMaster> ais523, åäö ? does those look correct
15:03:06 <ais523> I'm not sure what they're meant to look like, but I assume so
15:03:09 <ehird> it wasn't mojibake.
15:03:10 <AnMaster> anyway my client is set to auto detect
15:03:14 <ais523> a with a weird accent, a umlaut, o umlaut
15:03:17 <AnMaster> ais523, aao with ring, 2 dots, 2 dots
15:03:55 <AnMaster> ais523, and as wikipedia will tell you, at least in Swedish they are separate chars, not just variants of a and o
15:04:11 <AnMaster> think it is same in Norwegian except they use ø
15:05:57 <oerjan> AnMaster: yes, also the order is different
15:06:36 <ais523> anyway, as for that bug in Google I was talking about: http://groups.google.com/group/frc-play/topics
15:06:38 <oerjan> AnMaster: i think peng[ae]r <- penning[ae]r
15:07:08 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | :p.
15:07:15 <fizzie> The Finnish alphabet has the same "åäö" as the Swedish one, although å (the "Swedish O") is not used in any Finnish words, just names and stuff like that.
15:07:38 <AnMaster> oerjan, sounds like a weird form though, not something I would write/use. Probably oldish/poemish or something
15:07:58 <oerjan> AnMaster: it's an old coin unit
15:09:07 <oerjan> 1 riksdaler = 192 penningar
15:09:08 <ehird> ais523: the download link works for me
15:09:16 <oerjan> http://en.wikipedia.org/wiki/Swedish_riksdaler
15:09:31 <AnMaster> oerjan, blergh, Good thing we use a more logical system since ages
15:09:52 <ais523> ehird: yes, the point is it shouldn't be there in the first place
15:10:03 <Deewiant> Yes, because base-10 is more logical than any other base. (What?)
15:10:07 <ais523> the "read more" link shouldn't be anyway
15:10:12 <ais523> it puts ... at the end of the message
15:10:25 <fizzie> Yes, but the å letter is called "the swedish O" to distinguish it from the "o" letter. I'm not sure how much sense that makes.
15:10:35 <AnMaster> Deewiant, well yes it is, for a simple reason... Our system is base 10 elsewhere
15:10:52 <ehird> ais523: final newline, maybe
15:11:03 <Deewiant> AnMaster: what system? Times are base-60 or base-24. Dates vary.
15:11:19 <AnMaster> Deewiant, well the Babylonians used base 60 for "counting" iirc.
15:11:29 <ais523> ehird: no, because the second message in that thread is longer
15:11:32 <AnMaster> but we have 0-9 before we get a digit in the next position
15:11:32 <ais523> and also has the read more link
15:11:58 <AnMaster> Deewiant, thus I'd say we use base 10 (exception: programmers)
15:12:18 <AnMaster> Deewiant, we write 15 minutes, not F minutes
15:12:34 <Deewiant> AnMaster: well yeah, we write everything in base 10. That doesn't mean it's any more /logical/ to use base 10 for other things, especially retroactively.
15:12:34 <AnMaster> thus I'd say minutes are base-10 but modulo 60
15:13:18 <AnMaster> Deewiant, why isn't it a good idea to change to a more logic system?
15:13:29 <ais523> AnMaster: time is base-60, but each digit is written in base 10
15:13:34 <ais523> so it's decimal-coded-segasdecimal
15:13:49 <Deewiant> What is so much more logical about 10 than about 60
15:13:59 <Deewiant> It might be the more obvious choice for humans, sure
15:14:04 <Deewiant> But it is no more or less logical
15:14:15 <AnMaster> Deewiant, also why did you say retroactive?
15:14:38 <Deewiant> AnMaster: What do you mean, why?
15:14:55 <AnMaster> But did we change money retroactively?
15:15:06 <AnMaster> We replaced the current system with a new one
15:15:20 <AnMaster> but we didn't rewrite history to never had used the replaced system
15:16:18 <Deewiant> Hmm, I think I misunderstood myself
15:16:40 <Deewiant> Doing it retroactively is even worse :-P
15:16:41 <AnMaster> Deewiant, as for why base 10, I agree it isn't any more inherently logical than anything else. But for beings with 10 fingers, there is a certain logic yes
15:16:54 <Deewiant> AnMaster: "logic" is just the wrong word IMO.
15:16:56 <AnMaster> Deewiant, got an example of it being done retroactively?
15:17:24 <AnMaster> then that is purely hypothetical, at least until ais523 implements Feather
15:18:49 <oklocod> most things in humans come in two, and the only reasonable way to count with one's fingers is to use binary
15:20:41 <ehird> oklocod: i don't think whoever thought of base-10 could have counted in binary on their fingers.
15:20:45 <ehird> or at least, e could have
15:20:45 <AnMaster> oklocod, sure, two lungs, two <whatever the English word is for "njure">... But... one heart? one brain? one stomach? One nose? One mouth?
15:20:50 <ehird> but e wouldn't have intuitively thought of it...
15:21:10 <AnMaster> Deewiant, blergh yeah in English
15:21:16 <ehird> oklocod has 4.3 brains
15:21:22 <ehird> AnMaster: what is njure
15:21:23 <AnMaster> then it depends on the language of choice too
15:21:27 <Deewiant> well yeah, they're brain-halves in english as well
15:21:31 <Deewiant> but I'd say they're two brains
15:21:44 <Deewiant> in the same way that we have two lungs and not two lung-halves
15:21:46 <AnMaster> ehird, kind of related to the liver. Humans have two of them
15:22:27 <AnMaster> Deewiant, the lungs are more separate though.
15:22:40 <Deewiant> how so? They're less separate in that they're at least pretty much identical
15:22:44 <oklocod> AnMaster: heart does not exist for this purpose.
15:23:05 <AnMaster> well then doesn't the other ones either
15:23:29 <oklocod> if you're choosing your base based on the number of things in your body, you're probably not smart enough to see what's inside it.
15:23:41 <AnMaster> so lets see, one nose, one mouth, one torso, two arms, two legs, *ten fingers*, *ten toes*
15:24:19 <oklocod> now change the topic, base 10 rage is building up inside me
15:24:21 <AnMaster> oklocod, which is easier to count before you invented the mirror? your ears or your fingers?
15:24:37 <ais523> some cultures use base 5, IIRC
15:24:50 <AnMaster> ais523, yes makes sense for counting on one hand or so
15:24:52 <oklocod> yes sure sure base 5 and also 20
15:25:04 <AnMaster> didn't the babylonians use base 60?
15:25:09 <oklocod> and perhaps a cool 60 slipped in at some point
15:25:36 <oklocod> but they all sucked ass, binary an base -2i are the way to go
15:26:16 <AnMaster> No I suggest one of these bases: e, pi, 42
15:26:16 <ais523> oklocod: -2i? Can you express all complex integers in that?
15:26:30 <ais523> base -2 works for all decimal integers, with the digits 0 and 1
15:26:38 <AnMaster> the first two allows you to get a precise value for certain transcendental numbers
15:27:18 <Deewiant> all complex integers using only four digits and no sign
15:27:42 <GregorR> Since 'i' is discrete, base <anything>i can't represent every number.
15:27:46 <ais523> Deewiant: what digits do you need? I think either 0123 or 0 1 i 1+i would work
15:27:58 <ais523> GregorR: complex integer is what I was talking about
15:27:58 <Deewiant> http://en.wikipedia.org/wiki/Quater-imaginary_base
15:28:03 <AnMaster> GregorR, um, what about you cancel the base out?
15:28:26 <oklocod> GregorR: duh, you can never represent every number
15:28:58 <oerjan> i resent your present representation
15:31:27 <AnMaster> I would suggest not only changing the base, but more too
15:31:56 <AnMaster> lets use base e and make use of a logarithmic scale
15:34:21 * oerjan has been fond of hyperbolic tangent scale since he discovered you could add relativistic velocities with it
15:35:42 <ais523> anyway, I got two emails about building services at the moment, and I'm not sure whether to be pleased or worried
15:35:47 <ais523> one saying that my email was working again
15:35:56 <ais523> and the other saying that they'd fixed the Door properly this time...
15:37:56 <ehird> <oklocod> GregorR: duh, you can never represent every number
15:38:06 <oerjan> ais523: well the first one was apparently not entirely wrong, i guess?
15:38:23 <ais523> but every door message seems to have caused it to have gotten worse
15:38:30 <ais523> although the time it worked for me but nobody else was amusing
15:38:42 <ais523> (although annoying due to all the time I had to spend opening it for other people)
15:38:48 <oerjan> as for the second, i eagerly await when you will start climbing in the windows...
15:39:21 <ehird> oklocod: i like representing numbers.
15:39:29 <ehird> i vant perfect computeral arithmetic!
15:39:51 <oerjan> numbers demand representation!
15:39:53 <oklocod> computerolous arithmetology
15:39:54 -!- ais523_|direct has joined.
15:40:04 -!- ais523_|direct has left (?).
15:41:11 <oerjan> hm that implies someone once said something containing "washing the windows api"
15:42:01 <ais523> it was markov-chaining
15:42:09 <ehird> ais523: no, that was pretty literal
15:42:12 <ehird> just two sentences put together
15:42:15 <ais523> so "washing the windows" and "the windows api"
15:42:21 <ais523> fizzie found the sources for us
15:42:33 -!- Judofyr has quit (Read error: 104 (Connection reset by peer)).
15:42:34 <oerjan> i thought it based the next word on the 3 previous ones?
15:42:40 <ais523> on the two previous, I think
15:43:02 <ais523> fizzie: settle this argument for us?
15:43:09 -!- Judofyr has joined.
15:43:20 <ehird> order-2 = 2 previous ones
15:45:01 <AnMaster> hm lower order would give more non-sensical output?
15:45:57 <AnMaster> I guess order-0 wouldn't be a markov chain any more?
15:46:36 <ehird> order-0 would be totally random
15:46:40 <ehird> it'd pick more common words more
15:46:47 <ehird> so it'd just be "here's a random common word"
15:55:18 -!- Ilari has quit ("Won't be back for a while...").
15:56:25 <fizzie> It uses 4-grams, so it has a context of three words (in most cases) to choose the following word.
15:56:50 <ais523> hmm... as oerjan said, that would imply that "washing the windows api" was actually in a message somewhere
15:57:04 <ais523> unless it reduced to 3-grams for that because it couldn't find many 4-grams?
15:57:07 <fizzie> It might've been 4-grams when it outputted that.
15:57:19 <fizzie> I've been using various model orders throughout its history.
15:58:23 <oerjan> fungot, say it isn't so
15:58:23 <fungot> oerjan: fconv merely _returned_ 0; it didn't _print_ it.
15:58:45 <fungot> oerjan: blah i was all about and i didn't have experimental selected so it was discussed very shortly, then matthew announced the decision to go in 2 bits
15:59:10 <fizzie> And yes, it will also currently use a 3-gram if it can't find *any* 4-grams that have the current three-word context as their initial 3 words. I'm not sure if that should normally happen, since the three previous words will have been generated with the same model.
15:59:29 <fizzie> No-one seems to be washing the API in my logs; just windows in general.
16:05:55 -!- sebbu has joined.
16:23:02 -!- oerjan has quit ("leaving").
16:46:02 -!- Slereah_ has quit (Read error: 104 (Connection reset by peer)).
16:47:25 -!- Slereah_ has joined.
17:08:41 -!- oerjan has joined.
17:30:17 -!- Slereah has joined.
17:35:53 -!- KingOfKarlsruhe has quit (Remote closed the connection).
17:42:22 -!- Slereah_ has quit (Read error: 110 (Connection timed out)).
17:55:19 -!- LinuS has quit (Read error: 104 (Connection reset by peer)).
17:57:19 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | how goes "interfunge"?.
17:57:27 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | aaah.
17:57:29 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | I'm a programmer, not a lawyer, dangit!.
17:58:02 <oerjan> interfunge, what was that?
17:58:33 <ais523> Befunge written in INTERCAL
17:59:51 -!- olsner has joined.
18:03:29 <ehird> ais523: that was anmaster talking about ICAL actually
18:03:56 <ais523> because AnMaster likes using names thar were already taken, presumably
18:04:14 <ais523> also, ehird, do you mean IFFI?
18:04:21 <AnMaster> ICAL? That is a fingerprint that Mike Riley did
18:22:08 -!- KingOfKarlsruhe has joined.
18:29:16 -!- KingOfKarlsruhe has quit (Remote closed the connection).
18:42:49 <psygnisfive> did you know that theres a natural language formalism thats heavily based on types and lambdas?
18:43:17 <psygnisfive> theres even a version that depends on composition of functions, and on type raising
18:44:16 -!- LinuS has joined.
18:44:19 <oklocod> well i do know now, it's fairly irrelevant whether i knew before
18:44:26 <oklocod> because you can never prove i didn't
18:44:46 <psygnisfive> its called categorial grammar. the funky version is combinatory categorial grammar
18:45:54 <psygnisfive> some verbs have the type S\NP which means that they produce an S when they merge with an NP that's on their left
18:46:32 <psygnisfive> then John runs is S, because NP S\NP produces S, as S\NP states
18:47:27 <psygnisfive> for more arguments to the verb: bites :: (S\NP)/NP
18:47:52 <oklocod> well i find the notation fairly weird
18:48:24 <psygnisfive> the notation isnt that bad actually, given that order is relevant here
18:48:36 <psygnisfive> X\Y means you get an X if you left-merge with a Y
18:48:49 <oklocod> but i only got it just after calling it weird.
18:48:51 <psygnisfive> its kind of like an abbreviation of CFG rules
18:50:44 <psygnisfive> the sequence X\Y Y\Z Z can be analyzed as X\Y (Y\Z Z)
18:51:14 <psygnisfive> or you can do composition and get X\Z = (X\Y Y\Z); X\Z Z
18:51:24 <psygnisfive> which lets you handle all sorts of crazy discontinuities
18:53:43 <psygnisfive> sorry, had to get the slides to make sure i had the notation correct
18:54:15 <oklocod> you don't have to hurry, i stared at that for about a minute before realizing it was trivial.
18:54:49 <oklocod> (more than a minute, emphasis on not the realizing but the time it took)
18:55:46 <psygnisfive> so Marcel::S/(S\NP) ran::S\NP => Marcel ran:S
18:56:09 <psygnisfive> this seems pointlessly trivial but it makes it completely trivial then to handle sentences like
18:56:16 <psygnisfive> "Marcel proved and I disprove completeness"
18:57:31 <psygnisfive> Marcel:NP proved:(S\NP)/NP and:(X\X)/X I:NP disproved:(S\NP)/NP completeness:NP
18:58:12 <psygnisfive> Marcel:S/(S\NP) proved:(S\NP)/NP and:(X\X)/X I:S/(S\NP) disproved:(S\NP)/NP completeness:NP
18:58:22 <psygnisfive> function compose Marcel with proved, and I with disproved:
18:58:55 <oklocod> haha, congrats on the pun :P
18:59:31 <psygnisfive> [Marcel proved]:S/NP and:(X\X)/X [I disproved]:S/NP completeness:NP
19:00:25 <psygnisfive> [Marcel proved]:S/NP [and I disproved]:(S/NP)\(S/NP) completeness:NP
19:00:37 <psygnisfive> [Marcel proved and I disproved]:(S/NP) completeness:NP
19:00:41 <oklocod> btw: i dropped near "ok so"
19:00:50 <psygnisfive> [Marcel proved and I disproved completeness]:S
19:01:47 <oklocod> wait a mo, i'll try to understand all this.
19:03:02 <oklocod> okay i get it to some extent.
19:03:17 <psygnisfive> its just function composition, and type raising
19:04:04 <oklocod> i don't know what function composition is in this context
19:04:14 <Asztal> so.. Marcel:S/(S\NP) proved:(S\NP)/NP --> [Marcel proved]:S/NP
19:04:23 <oklocod> why can NP become S/(S\NP)?
19:04:30 <Asztal> it's kind of intuitive
19:04:59 <oklocod> S/(S\X) is, of coutse, just X
19:05:19 <oklocod> because it's S with something on its right, that has S on its left
19:05:28 <psygnisfive> you get an S when you merge on the right with something that needs an X on the left to make an S
19:05:42 <psygnisfive> something that needs an X on its left to MAKE an S
19:06:01 <psygnisfive> f x is prefix notation for applying f to x right?
19:06:14 <psygnisfive> but why cant it be postfix notation for calling method f on x?
19:06:35 <psygnisfive> who says x is the argument and f is the function?
19:06:43 <psygnisfive> why cant x be the function and f the argument?
19:06:58 <psygnisfive> i mean, numbers can be modelled as functions right?
19:07:35 <psygnisfive> and binary functions over booleans are often modeled in LC as using the BOOLEANS as the functions, right?
19:08:45 <oklocod> i'd have booleans be a universal operation like nor, but yeah go on
19:09:17 <oklocod> so, marchel is a noun thingie
19:09:34 <oklocod> when you do Marcel:S\(S/NP)
19:10:08 <psygnisfive> type raising, but we did S/(S\NP) since marcel was the subject of the verb :p
19:10:49 <oklocod> something that takes an S on the <apply direction> to produce something that takes an S on the <other direction> to produce an NP
19:11:06 <psygnisfive> X/Y means "produce an X by taking a Y on the right
19:11:13 <oklocod> i'm refusing to believe it's not what i originally thought it is.
19:11:29 <psygnisfive> so S/(S\NP) says "produce an S by taking an S\NP on the right"
19:11:56 <oklocod> although i think these are the same concept
19:12:22 <psygnisfive> well you might be able to say that instead of this:
19:13:28 <oklocod> yes yes it's all clear now
19:13:41 <psygnisfive> you can derive whole sentences that way dude
19:14:00 <psygnisfive> without types: I believe that she ate dinner
19:14:39 <psygnisfive> with types (after some type raising): I:S/VP believe:VP/S' that:S'/S she:S/VP ate:VP/NP dinner:NP
19:14:45 <psygnisfive> well thats a nice function composition chain there
19:14:57 <psygnisfive> [I believe]:S/S' that:S'/S she:S/VP ate:VP/NP dinner:NP
19:15:04 <psygnisfive> [I believe that]:S/S she:S/VP ate:VP/NP dinner:NP
19:15:19 <psygnisfive> [I believe that she]:S/VP ate:VP/NP dinner:NP
19:15:55 <psygnisfive> and thats like, COMPLETELY opposite of how most syntactic models look at sentence structure
19:16:12 <lament> type theory for sentences?
19:16:32 <psygnisfive> also, theres a formalism for how to equate these things with their truth conditions
19:16:47 <psygnisfive> so that these applications also can produce lambdas and such
19:16:48 <lament> how is it the opposite of syntactic models?
19:17:11 <psygnisfive> the normal idea about the structure of that last sentence would be more like
19:17:45 <psygnisfive> that + [she ate dinner] -> [that she ate dinner]
19:17:55 <psygnisfive> believe + [that she ate dinner] -> [believe that she ate dinner]
19:18:15 <psygnisfive> I + [believe that she ate dinner] -> [I believe that she ate dinner]
19:18:15 <olsner> and then you also have monadic discourse models :)
19:18:31 <oerjan> AAAAAAAAAAAAAAAAAAAAAAAAA
19:19:06 <psygnisfive> give1: ((S\NP)/NP)/NP : \z\y\z[give(x,y,z)]
19:19:49 <psygnisfive> so each left or right merge applies the lambda
19:20:39 <oklocod> that might be nice in a language
19:21:03 <AnMaster> psygnisfive, also did you see that idea I had for a "HTML Query Language"?
19:21:03 <psygnisfive> english is pretty esoteric, ill have you know
19:21:36 <AnMaster> INSERT ELEMENTS head, body INTO THE ELEMENT head;
19:21:37 <oklocod> the cases were so up you could see them from china.
19:21:47 <AnMaster> lots of statements like that to create a HTML document
19:21:58 <AnMaster> psygnisfive, instead of that messy <html> stuff
19:22:37 <AnMaster> okt 19 16:10:48 <AnMaster> SELECT TEXT OF ELEMENT p WHEN ATTRIBUTE id OF ELEMENT p IS EQUAL TO TEXT VALUE "top" AND ALSO TEXT OF ELEMENT p STARTS WITH TEXT VALUE "ehird";
19:22:39 <Asztal> E4X probably does half of it :)
19:23:00 <psygnisfive> but less so than doing the same with JS, probably
19:23:13 <AnMaster> okt 19 16:22:13 <AnMaster> UPDATE TEXT OF THE FIRST ELEMENT p OF ELEMENT body OF THE ELEMENT html SETTING NEW VALUE TO TEXT "Blergh...";
19:23:16 <olsner> html already has a query language: css
19:23:20 <lament> ITS QUITE IMPORTANT THAT YOUR LANGUAGE USES ALL CAPITALS, THAT WAY PEOPLE KNOW IT'S A REAL QUERY LANGUAGE AND NOT SOME FAKE CRAP, BECAUSE EVERYONE KNOWS REAL QUERY LANGUAGES ALWAYS USE UPPERCASE
19:23:43 <AnMaster> okt 19 16:24:31 <oerjan> "This language is insensitive to everything, except case"
19:23:45 <Asztal> well, with E4X (a javascript extension) you can create XML literals and query them to your heart's content
19:24:00 <olsner> well, not directly, but it has an embedded query language for specifying which elements a style applies to
19:24:09 <ehird> lament: IT'S QUITE IMPORTANT THAT YOU RECOGNIZE THAT IT IS A JOKE, ALBEIT QUITE A BAD ONE
19:24:30 <psygnisfive> you should also require that it be in lolcatese
19:24:47 <AnMaster> psygnisfive, also oerjan suggested that each statement should end with ", OR ELSE!" as a opposite of INTERCAL's "PLEASE"
19:25:33 <AnMaster> okt 19 16:02:08 <AnMaster> INSERT ELEMENTS head, body INTO ELEMENT html; INSERT ELEMENT title INTO ELEMENT head OF ELEMENT html; INSERT TEXT "this is a horrible idea for markup" INTO ELEMENT title OF ELEMENT head OF ELEMENT html;
19:25:52 <AnMaster> though I later decided that the top element needs "THE"
19:26:02 <psygnisfive> I CAN PUT ELEMENTS head, body IN ELEMENT html
19:26:17 <AnMaster> psygnisfive, I'm not interested in a lolcat version
19:26:46 <AnMaster> psygnisfive, it haven't been speced, so how could you fork it?
19:27:06 <AnMaster> psygnisfive, since specing this may lead to insanity
19:27:10 <oerjan> you will not be allowed to fork it! we'll make it closed source! and patent it!
19:27:38 <AnMaster> psygnisfive, just I think lolcat doesn't add anything to the joke
19:27:47 <AnMaster> psygnisfive, rather it should look more like COBOL
19:28:08 <olsner> lolcatting doesn't add much if anything to an all-caps cobolish language
19:28:10 <psygnisfive> so instead of fork::lang/lang, i typeraised your language to lang\(lang/lang)
19:28:26 <AnMaster> psygnisfive, you make no sense.
19:28:37 <olsner> sense, you makes none!
19:28:40 <psygnisfive> you just didnt read anything i said about combinatory categorial grammars :P
19:28:57 <AnMaster> psygnisfive, nor do I intend to, natural languages are boring most of the time
19:29:00 <Asztal> speaking of which, how would you parse "The horse raced past the barn fell."? Backtracking?
19:29:00 * oerjan looks at oklocod sitting catatonic in the corner
19:29:08 <psygnisfive> yeah but dude this isnt just natural language
19:29:20 <psygnisfive> this is function composition and currying and type raising!
19:29:22 <olsner> psygnisfive: requiring knowledge of combinatory categorial grammars is basically equivalent to not making sense :)
19:29:53 <oerjan> <psygnisfive> olsner, we're in #esoteric <psygnisfive> be serious
19:30:00 <oerjan> i sense a cognitive dissonance
19:30:03 <lament> Asztal: i don't know much about grammar theory but intuitively that seems as complex as regexes
19:30:15 <olsner> well, true... just *being* here is making no sense
19:30:45 <olsner> oerjan: cognitive dissonance is definitely on-topic here
19:31:09 <olsner> T-something attribute grammars?
19:31:09 <psygnisfive> tree-rewriting models for natural language syntax
19:31:24 <oklocod> psygnisfive: algorithmmmmms
19:31:53 <olsner> now that I think about it, I was thinking about modelling a card game on a three-adjoining grammar at one point
19:32:13 <oklocod> i love this book, just algorithm after another
19:32:25 -!- oklocod has changed nick to oklopol.
19:32:45 <oklopol> algorithm design, jon kleinberg & eva tardos
19:33:11 <psygnisfive> well, this, but she practically says these
19:33:19 <AnMaster> <<<<(X*2)>>||<<X>><=<<1,2,3>>>>.
19:33:30 <oklopol> can't get it over the internets afaik, so you won't do much with the name
19:33:39 <AnMaster> psygnisfive, does your stuff help you understand that?
19:34:36 <oklopol> and you cannot look at nopol, i don't publish my languages except in this channel, when they are born :P
19:34:46 <AnMaster> psygnisfive, that<AnMaster> <<<<(X*2)>>||<<X>><=<<1,2,3>>>>.
19:34:54 * oklopol will put noprob on the wiki if the interpreter ever finishes, though
19:35:08 <AnMaster> psygnisfive, I bet your weird natural languages models doesn't help you in understanding that :P
19:35:14 <oklopol> psygnisfive: yeah, that's nopol
19:35:18 <oerjan> oklopol: sounds like a halting problem to me
19:35:34 <psygnisfive> so it doesnt. but that doesnt mean CCGs couldnt!
19:35:41 <oklopol> oerjan: i was afraid i might trigger a joke :P
19:35:42 <AnMaster> psygnisfive, it's actually not an esolang.
19:35:55 <AnMaster> psygnisfive, it is a functional language, and using some very weird syntax from it
19:36:46 <psygnisfive> the main difference between natlangs and complangs is that complangs generally have very shallow semantics and very clear structure, so its not hard to talk about them
19:37:00 <AnMaster> oerjan, yep, and using "bit string comprehensions"
19:37:01 <psygnisfive> the type of something in a programming language is purely a matter of value
19:37:05 <AnMaster> but removing all the usual whitespaces
19:37:15 <AnMaster> << << (X*2) >> || <<X>> <= << 1,2,3 >> >>.
19:37:20 <AnMaster> would be the normal way to write it
19:37:28 <psygnisfive> but in natural language we have lots of shit to do with not just "Value" but also with representation type
19:37:53 <AnMaster> oerjan, same concept as list comprehensions
19:37:59 <psygnisfive> because things can be represented in various ways
19:38:11 <psygnisfive> i mean, just consider what makes a noun a noun
19:38:12 <AnMaster> psygnisfive, ever heard of Feather?
19:38:33 <AnMaster> psygnisfive, well, that isn't shallow I think
19:38:36 <Asztal> a miserable little pile of semantics!
19:38:42 <AnMaster> it is retroactively non-shallow
19:38:58 <oklopol> psygnisfive: so, how about making a language that has such a complicated and exception-ridden syntax no one will ever be able to write it, it could be based on your wonky syntactic theories
19:38:59 <AnMaster> retroactive changes to the own grammar rocks
19:39:17 <psygnisfive> oklopol: ive been saying we should for months now :P
19:40:31 <oklopol> basically, we start with some simple structure, and start building incredibly complicated sublanguages, and add exception on top of exceptions until it's a total mess, after which we start cleaning it up, making it *look* simple, in short and simple programs
19:40:36 <oklopol> but the underlying semantics
19:40:55 <oklopol> i dunno, but yeah, i guess i wanna do something like that
19:41:04 <psygnisfive> we'd also have to have atleast two ways of representing the same thing
19:41:23 <oklopol> BUT, i've read 12 pages today (two exams, and i slept like 3 hours during the day), my quota is more like 80
19:41:27 <psygnisfive> similar to how you can talk about events in english using sentences, or using noun phrases
19:41:45 <oklopol> perl doesn't have a complicated grammar
19:41:56 <AnMaster> "<psygnisfive> we'd also have to have atleast two ways of representing the same thing <psygnisfive> each with its own quirks of distribution"
19:42:08 -!- Judofyr has quit (Remote closed the connection).
19:42:23 <oklopol> well we're gonna make perl^7-2
19:42:25 <lament> there is more than two ways to do it.
19:42:41 <oklopol> yeah probably more like five
19:42:48 -!- Judofyr has joined.
19:43:21 <psygnisfive> yes but are they describable with type raising, function composition, and so on?
19:43:29 <psygnisfive> because thats how our language will be described
19:44:25 <psygnisfive> i say we use our language to compile down into simple predicates
19:44:27 <AnMaster> <Asztal> speaking of which, how would you parse "The horse raced past the barn fell."? Backtracking?
19:44:31 <AnMaster> psygnisfive, did you answer that?
19:44:33 <psygnisfive> and then run it on top of those predicates
19:45:09 <AnMaster> psygnisfive, just answer Asztal's question
19:46:08 <psygnisfive> i mean, well let me rephrase that since you're not really asking about formalisms i guess
19:46:22 <psygnisfive> i couldnt tell you how a PARSER would work on that, for two reasons
19:46:43 <psygnisfive> 1) the formalism im familiar with is notoriously hard to parse, supposedly, and i've never worked on a parser for it
19:47:03 <psygnisfive> 2) CCG formalisms i dont know much about, nevermind CCG parsers
19:47:13 <psygnisfive> tho i can link you to a paper on parsing with CCGs
19:48:06 * oerjan gets a barn fell to race a horse past
19:48:43 * oerjan swats psygnisfive -- er wait no
19:49:00 * oerjan clobbers psygnisfive with a hammer
19:49:33 <AnMaster> the horse raced past the barn (the barn fell)
19:49:44 <oerjan> Anyone want to buy a barn fell, cheap?
19:50:01 <psygnisfive> theres only one valid parse for that sentence, its just garden pathy
19:50:04 <AnMaster> the horse, raced past the barn, fell
19:50:19 <psygnisfive> ok ok listen guys thats not what it means :p
19:50:27 <AnMaster> psygnisfive, http://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffalo_buffalo_buffalo_Buffalo_buffalo
19:50:29 <psygnisfive> the horse that was raced past the barn fell down
19:50:36 <psygnisfive> anmaster: i can get more buffalo than that, actually
19:51:03 <AnMaster> "The horse (that was raced past the barn) fell."
19:51:07 <AnMaster> http://en.wikipedia.org/wiki/Garden_path_sentence
19:51:22 <AnMaster> <AnMaster> the horse, raced past the barn, fell
19:51:22 <AnMaster> <psygnisfive> ok ok listen guys thats not what it means :p
19:51:41 <psygnisfive> commas are used for specific things in english
19:51:47 <psygnisfive> they dont denote relative clauses like in your language
19:52:07 <AnMaster> psygnisfive, why do you think they do in [my language]?
19:52:13 <psygnisfive> "the horse raced past the barn of its own accord), and then fell"
19:52:42 * oklopol remembers a horce raced past
19:52:44 <psygnisfive> i presume you're some sort of finnogermanic like half the rest of #esoteric :P
19:52:46 <AnMaster> psygnisfive, not valid, unmatched parentheses
19:53:17 <AnMaster> psygnisfive, what is that then?
19:54:09 <AnMaster> psygnisfive, My native language does not have "ü"
19:55:03 <oerjan> oklopol: what does it mean? :D
19:55:16 <oklopol> well not sure, but it has "yy", and "ä"
19:55:22 <AnMaster> psygnisfive, I don't speak any of the languages you tried
19:55:36 <AnMaster> psygnisfive, So why the insult that I'm some sort of food?
19:55:59 * AnMaster gives psygnisfive a bad stomach
19:56:25 <AnMaster> psygnisfive, and I just ate a lot of garlic.
19:56:36 <oerjan> is it furry curry, in a hurry?
19:57:03 <AnMaster> psygnisfive, I'm from Sweden. I don't speak any of the languages you gussed
19:57:05 <psygnisfive> take a whole head of garlic and remove the papery outer crap
19:57:17 <AnMaster> and I still don't get what "finnogermanic" means when used about a person
19:57:21 <AnMaster> "<psygnisfive> i presume you're some sort of finnogermanic like half the rest of #esoteric :P"
19:57:32 <psygnisfive> step 3: coat with olive oil and sprinkle with oregano
19:57:42 <psygnisfive> step 4: bake for 45 minutes to an hour at 350 to 400 *F
19:57:50 <oerjan> AnMaster: it means you speak a language of the finnogermanic family *ducks*
19:57:57 <psygnisfive> step 5: remove, let cool till warm, then up turn and squeeze the sides
19:58:03 <AnMaster> oerjan, that still doesn't explain anything
19:58:06 <psygnisfive> Anmaster: swedish is a north germanic language
19:58:17 <AnMaster> psygnisfive, yes and? None of your guesses were correct on Swedish
19:58:19 <psygnisfive> germanic languages use commas differently than in english, usually for relative clauses
19:58:36 <psygnisfive> hence why i commented that you're finnogermanic
19:58:41 <oerjan> psygnisfive: danish uses lots of commas, norwegian not that much
19:58:54 <AnMaster> oerjan, Swedish doesn't use much I think, I guess it is relative though
19:58:57 <psygnisfive> i know in german atleast commas are relative clauses
19:59:12 <psygnisfive> the boy, that i fucked like a bitch, is named dylan
19:59:28 <psygnisfive> whereas in english thats completely invalid use of commas
19:59:30 <AnMaster> psygnisfive, No pedophiles please
19:59:50 <AnMaster> psygnisfive!*@* added to ignore list.
20:00:27 <olsner> hmm, that's probably in the grey area between pedo- and ephebophilia
20:00:31 <oklopol> oerjan: is it furry curry, in a hurry? <<< furry doesn't rhyme here
20:00:52 <AnMaster> oklopol, depends on which furry I guess
20:01:07 <ehird> AnMaster seems to have some kind of phobia of any reference to pedophillia at all
20:01:10 <olsner> actually, according to wikipedia, that's clearly pedophilia rather than ephebophilia "Ephebophilia refers to the sexual preference for adolescents around 15-19 years of age."
20:01:55 <olsner> maybe hebephilia rather than pedophilia though
20:01:58 <oerjan> exactly who was going around joking about tusho rape some while ago...
20:02:08 <oklopol> i had a 13-yo gf about a year ago
20:02:48 <olsner> let me rephrase that as "13-yo gf == tusho?"
20:02:50 <oklopol> dunno, i guess it's illegal
20:02:51 <ehird> AnMaster: you DIDN'T ignore him
20:03:11 <psygnisfive> tusho is distinctly male, despite the humor of saying he's female
20:03:13 <ehird> AnMaster: no you didn't , you just confirmed one of his statements
20:03:20 <AnMaster> ehird, I just tried to respond to olsner's questions
20:03:36 <AnMaster> <oklopol> i had a 13-yo gf about a year ago
20:03:36 <AnMaster> <AnMaster> olsner, tusho == ehird yes
20:03:48 <psygnisfive> everyone knows that when you ignore someone it doesnt get announced to the world!
20:03:58 <AnMaster> since he seemed confused I thought I'd explain it
20:04:04 <olsner> well, it does if you quote the message you got from your client
20:04:30 <AnMaster> olsner, now that made no sense heh
20:04:36 <psygnisfive> see anmaster, thats what happens when you block people
20:04:37 <olsner> that was definitely a normal message string
20:04:49 <oklopol> AnMaster: did i get ignored? wouldn't it be kinda weird to ignore someone for a joke, and not for an actual crime :\
20:05:07 <AnMaster> oklopol, I didn't ignore you oklopol
20:05:23 <oklopol> psygnisfive: yes, that was my point
20:05:28 <AnMaster> oklopol, but you were clearly joking. While psygnisfive seemed serious
20:05:29 <psygnisfive> actually he ignored me for making fun of him using a joke that he set up in the first place
20:05:29 <ehird> AnMaster very often says "please no pedophillia" or basically the same wording all the time
20:05:37 <oklopol> AnMaster: err, i was not joking
20:05:38 <ehird> AnMaster: how the fsck did he seem serious
20:05:52 <ehird> also, how on earth is /ignore the correct reaction to the rape of a child...?
20:05:53 <oklopol> while psygnisfive was clearly joking.
20:05:56 <psygnisfive> ehird: he doesnt realize that you and i havent actually consummated our love
20:06:02 <ehird> psygnisfive: quite
20:06:14 <AnMaster> ehird, it isn't, but I agree with ais523's reasons too
20:06:26 <AnMaster> I refer to his reasons for ignoring psygnisfive
20:06:36 <olsner> oerjan: your mind is weak
20:06:50 <ehird> "ais523's reasons"?
20:06:55 <psygnisfive> knowing that i prefer cock in my ass than my cock in someone elses ass is TMI?
20:07:01 <psygnisfive> i mean, its implicit in the fact that im gay
20:07:09 <psygnisfive> you had BOTH possibilities in your mind before!
20:07:19 <psygnisfive> i'd say that's a reduction of information, sir
20:07:27 <oklopol> everyone knows you're a bottom
20:07:37 <oklopol> you don't talk about anything else
20:07:39 <olsner> I generally assumed gay men could be both tops and bottoms
20:07:49 <olsner> why would it be implicit that you're a bottom?
20:08:10 <psygnisfive> olsner: no i meant it was implicit that i was either a top or a bottom (or a switch)
20:08:29 <psygnisfive> so by confirming one, im not actually providing MORE information than already provided by the knowledge that im gay
20:08:38 -!- fungot has quit (Read error: 131 (Connection reset by peer)).
20:08:46 <psygnisfive> therefore "TMI" is clearly illogical, because im actually ruling out, and thus removing, alternatives
20:08:58 <oklopol> that's what information is
20:08:59 <psygnisfive> so it cant be too MUCH information, since the result is that theres less!
20:08:59 <olsner> oh, so when you said you were a bottom you weren't saying that you were not a top?
20:09:10 <AnMaster> fizzie, seems like fungot crashed or such
20:09:28 <olsner> but, then you're ruling out possibilities, and thus providing information
20:09:38 <olsner> too much of which would be too much
20:09:42 <psygnisfive> yes but thats not the information he (you?) meant
20:09:54 <fizzie> AnMaster: Yes, I tried to do "^code 000f-p" to clear the ignoration counter (talking to it in a query) but for some reason it hung up. Might be some sort of a cfunge incompatibility, actually.
20:09:55 <olsner> me? he? you? I don't know!
20:10:01 <psygnisfive> namely, it was implied that the information was was too much was the idea of someones cock in my ass
20:10:04 -!- fungot has joined.
20:10:42 <olsner> ah, yes, oerjan's mention of information distracted me from the original issue
20:10:42 <psygnisfive> surely noone would care about the YES/NO of such things, in this scenario, but rather the actual content
20:10:48 <AnMaster> fizzie, well that makes no sense, g and p are simple and easy
20:10:58 <psygnisfive> but theres cock in ass in all three situations, top, bottom, or switch!
20:10:59 <oerjan> olsner: there was an original issue?
20:11:12 <AnMaster> fizzie, also did cfunge itself crash or just fungot?
20:11:12 <fungot> AnMaster: 21:01 bonjovn4 shit and stuff. have fun!
20:11:25 <olsner> well, explicit mention of "cock in ass" is usually considered TMI
20:11:27 <fizzie> AnMaster: Just fungot, of course. It might've depended on some RC/Funge UNDEF thing.
20:11:27 <fungot> fizzie: it'd take a while
20:11:40 <oerjan> psygnisfive: do you EVER refrain from quibbling whenever possible?
20:11:42 <olsner> but just the bottom/top distinction shouldn't be
20:11:43 <AnMaster> fizzie, and if you can reproduce it, rebuild with DEBUG build
20:11:54 <psygnisfive> oerjan: this is #esoteric. how could i do such a thing
20:12:10 <fizzie> AnMaster: ^code is implemented by appending "0R" to the input, sticking it into some place of fungespace, loading SUBR and executing a C there.
20:12:25 <oerjan> psygnisfive: you ruin half my jokes by explaining them...
20:12:46 <AnMaster> fizzie, can't think of any reason that would crash
20:12:56 <oerjan> (ok maybe TMI wasn't _entirely_ a joke)
20:12:58 <olsner> well, implicit is still implicit... and bottom/top could very well have referred to submissive/dominant personality traits rather than sexual practice
20:12:58 <fizzie> I'll try it with some tracing.
20:13:22 <AnMaster> fizzie, could you give a trace of what 1) happens 2) you think should happen instead along with 3) a 4 page description of why ;)
20:13:43 <psygnisfive> now granted, they tend to go together quite frequently
20:13:54 <psygnisfive> but in straight BDSM its quite common to have femdom
20:14:03 <psygnisfive> which is almost always a case of a bottom dom
20:14:09 <ehird> psygnisfive: This is quite irrelevant for #esoteric.
20:14:10 <psygnisfive> unless the woman has a strapon or something
20:14:12 -!- ais523 has left (?).
20:14:15 <ehird> Move it to #psygnisfives-sexual-ramblings or something.
20:14:30 <olsner> yes, let's abort this while we still can :)
20:14:35 <AnMaster> ehird, you finally agree with me?
20:14:46 <ehird> AnMaster: No. psygnisfive: I can blame it on your continuous rambling.
20:14:53 <olsner> interesting to hear about the finer distinctions though
20:14:55 <ehird> About utterly irrelevant stuff that nobody here cares about.
20:15:01 <ehird> olsner: There's always /msg.
20:15:08 <psygnisfive> true, but we wouldn't've gotten here if anmaster hadn't turned us into a tangent
20:15:36 <olsner> the tangent, while tangential, was still #esoteric's tangent
20:15:37 <AnMaster> fizzie, wait SUBR may be relative storage offset differently than for RC/Funge. that is all I can think of
20:15:40 <psygnisfive> i merely used a sentence as an example, but no, he had to go and act like i was talking
20:15:56 <psygnisfive> anmaster doesnt know about use/reference distinctions i think :(
20:15:58 <AnMaster> fizzie, I remember having to mess with that because Deewiant thought it should have been and so on
20:16:40 * oklopol cares about weird sex stuff
20:16:45 <psygnisfive> who asked me about parsing the garden path sentence?
20:16:52 <ehird> psygnisfive: I still don't understand why AnMaster ignored you for one sentence that was clearly a joke.
20:17:12 <oklopol> indeed, and wouldn't do it to me for a *non* joke
20:17:28 <psygnisfive> oklopol, how were you dating a 13 year old
20:17:29 * oklopol doesn't like being called a liar
20:17:32 <ehird> oklopol: pedophillia is OK if you don't talk about sex, duh. Now. On to more interesting things
20:17:47 <Deewiant> yes, having sex is fine as long as you don't talk about it
20:17:49 <psygnisfive> asztal: you asked about the parsing right?
20:17:50 <oklopol> ehird: that's better than "you're clearly joking"
20:18:12 <ehird> Deewiant: Yes, the real menace is referencing having sex with underaged peopple.
20:18:20 <psygnisfive> ok. well, i can only comment about minimalism and parsing
20:18:20 <fizzie> AnMaster: It seems to jump to the right place, execute "000f-p" just fine, but then it hits a 0 and reflects. Seems I've tried to use "A" to append, except that after loading SUBR the A instruction is SUBR's "set absolute mode".
20:18:32 <psygnisfive> namely, movement based syntax seems hard to parse, but there might be some ways
20:18:48 <fizzie> AnMaster: So now I just wonder why it used to work. Are A/O new things?
20:18:51 <AnMaster> fizzie, yes it is, the fingerprint was ret-conned by Deewiant and Mike Riley
20:19:03 <AnMaster> fizzie, it would have worked a few weeks back
20:19:26 <AnMaster> fizzie, blame Deewiant for breaking existing apps, which is what I warned would happen
20:19:32 <AnMaster> I take no responsibility for that
20:19:39 <Asztal> psygnisfive: I'm also curious what happens if the word order is free (e.g. Hungarian, sort of)
20:19:45 <ehird> AnMaster: OMG FIZZIE WILL HAVE TO CHANGE A PROGRAM A LITTLE BIT! HOW DARE THEY IMPROVE THINGS
20:19:45 <fizzie> Oh well, the fix is trivial.
20:19:50 <AnMaster> Deewiant, see!? I predicted that would happen.
20:20:04 <psygnisfive> minimalist approaches take such things to be something called scrambling
20:20:07 <ehird> Deewiant: betting he has me on ignore
20:20:15 <psygnisfive> which is a fancy way of saying "shit aint in the order we expect it to be! :("
20:20:27 <AnMaster> Deewiant, versioned fingerprints. And what ehird said is irrelevant
20:20:43 <ehird> versioned - mm i love the smell of useless bloat in the morning
20:20:44 <ehird> it smells like failure.
20:20:50 <ehird> and what i said is very relevant
20:21:25 <psygnisfive> free word order is generally taken to be the result of movement
20:21:34 <AnMaster> Deewiant, And never change existing, unless you reserve instruction/other value for a parameter for future use
20:21:57 <AnMaster> or just make a new fingerprint
20:24:57 -!- ehird_ has joined.
20:24:58 <ehird> AnMaster: stop bullshitting
20:25:00 -!- ehird__ has joined.
20:25:05 <ehird> as fizzie said - the change was trivial
20:25:11 <ehird> and it improves the fingerprint
20:25:16 <ehird> what you suggest only adds to bloat for no real gain
20:25:17 <AnMaster> huh, why so many clients of yours?
20:25:42 <AnMaster> I still hold the same opinion.
20:25:54 <ehird_> It'd be nice if you offered a real justification, but I know better than that.
20:26:40 <AnMaster> ehird_, for a simple reason: Not breaking existing code.
20:27:39 <AnMaster> Why do you think the C standard committee avoids breaking changes when possible? Why do you think old functions in both the POSIX standard and on Windows remains?
20:27:53 <AnMaster> And they make new ones if the old ones can't be upgraded easily
20:27:59 <ehird_> AnMaster: BECAUSE THAT IS C
20:28:10 <AnMaster> ehird, same goes for many other languages.
20:28:10 <ehird_> C IS USED FUCKING. EVERYWHERE. MISSION CRITICAL SYSTEMS DEPEND ON IT.
20:28:17 <ehird_> BEFUNGE IS A FUCKING ESOLANG
20:28:24 <ehird_> IN A _FUCKING IRC BOT_
20:28:43 <ehird_> yes, there is, because you have a continual and constant failure of basic logic and reasonability
20:29:17 <AnMaster> And well why do you think there are no mission critical systems in Befunge? Apart from it being a language where programs are hard to maintain, slow and so on?
20:29:35 <ehird_> AnMaster: maybe because it's an esolang that is slow, whose programs are hard to maintain?
20:29:39 <ehird_> maybe because it's AN ESOLANG
20:29:46 <AnMaster> ehird, yes but apart from that :)
20:29:54 <ehird_> ha ha ha you're making a joke to justify your idiocy
20:29:57 <ehird_> very funny, but it's not valid
20:30:35 <AnMaster> and I still stand by my point, breaking changes lead to bitrot
20:30:47 <AnMaster> in a language hard to maintain this is even worse
20:31:11 <Deewiant> I never questioned that; I just said that such programs would be written by people in this channel. fizzie hasn't disproved that yet.
20:31:15 <ehird_> AnMaster: To follow in the vein of your beaurocracy, please compile a list of programs that have been broken by the change, and your assesment of how hard it will be to fix them.
20:31:24 <ehird_> Once you can, then I will concede.
20:31:37 <ehird_> If you cannot, then I will continue to call your logic retarded andy our point invalid.
20:33:17 <AnMaster> ehird, I'm not omniscient, I can't know everything, if I were, such a change would be trivial, since I would be able to tell all affected.
20:33:37 <ehird_> AnMaster: Occams razor dictates that the change is fine.
20:34:07 <AnMaster> ehird, I don't see how you mean.
20:34:37 <ehird_> AnMaster: Considering your failure at logic I am not surprised.
20:34:38 <AnMaster> Deewiant, also well what about other places? Not everyone is here, for example Mike Riley often isn't
20:35:06 <AnMaster> ehird, well I do know what Occams razor is, I don't see what it has to do with breaking changes
20:35:18 <AnMaster> ehird, so unless you can justify that?
20:35:42 <ehird_> AnMaster: Generally trivial logic that a 3-year-old could understand does not need justification. Your warped sense of reality, however, does.
20:35:46 <Deewiant> AnMaster: and has he written a program broken by this change? Perhaps, but he was of course aware of the change. :-P
20:35:59 <AnMaster> Deewiant, yes, but that shows there may be other ones
20:36:11 <Deewiant> my point was that there likely aren't
20:36:16 <AnMaster> ehird, well then you don't seem to have anything valid to add, if you refuse to explain yourself
20:37:01 <ehird_> AnMaster: Trivial, undeniable, simple logic does not need justification. Your fucked-up justifications that turn out to actually be illogical do.
20:37:17 <AnMaster> ehird, I'm still waiting for your justification. I have given my reasons.
20:37:36 <ehird_> AnMaster: I have adequately explained why only an idiot would need justification of a basic application of Occam's Razor.
20:37:41 <AnMaster> ehird, I'm still waiting for your justification. I have given my reasons.
20:37:42 <ehird_> Do you want me to treat you as an idiot? I can if you want.
20:38:04 <AnMaster> It can still break programs. It can be hard to debug.
20:38:13 <AnMaster> Befunge programs are notorious for that.
20:38:19 <ehird_> Please read up. Thanks.
20:39:37 <AnMaster> ehird, and I'm waiting for justification. Yes you prefer the simple solution. But I argue breaking existing programs that complies to the then published specs is not valid
20:39:50 <ehird_> There. Is. Nothing. To. Justify.
20:39:53 <AnMaster> why do you think there is Funge-98 and Befunge-93?
20:40:04 <AnMaster> Why not just rewrite "Befunge" as it was then known
20:40:17 <ehird_> Jesus christ you are an idiot. Now I have to put you on /ignore in all of 3 clients.
20:52:20 <oerjan> fungot: at least _you_ are reasonable, right?
20:52:21 <fungot> oerjan: how's the scheme workshop of 2002, if i'm having connection issues, how would you answer ' which would be exactly that.
20:52:36 <oerjan> fungot: i'm sorry, i wasn't there
20:52:36 <fungot> oerjan: the cygwin page does say programs need to work much better.) originally developed by miller puckette and company at fnord.
20:53:54 <Deewiant> fungot: the guys at fnord made cygwin?!
20:53:54 <fungot> Deewiant: that's using generators. for streams see srfi-40 or http://xmog.com/ scrap/ show/ fnord
20:54:15 <Deewiant> fungot: xmog.com doesn't look like cygwin
20:54:15 <fungot> Deewiant: a lisp in php
20:54:31 <fungot> Deewiant: is there a special compiler i'll need, or to print to the channel
20:55:04 <Deewiant> fungot: for that lisp in PHP? Probably yes
20:55:05 <fungot> Deewiant: making your code `portable' only by virtue of the discussion flatt changed some of my init.el" at http://paste.lisp.org/ display/ 274
20:55:44 <fizzie> fungot: You're quite the lisper these days.
20:55:45 <fungot> fizzie: don't encourage him. it was created
20:55:58 <fizzie> fungot: Sounds suspiciously religious.
20:55:58 <fungot> fizzie: i can't make out what you mean
20:56:09 <Deewiant> fungot: that's some ugly lisp there
20:56:09 <fungot> Deewiant: painful i/ o
20:56:33 <fungot> oerjan: it's the cryptogram type puzzle fairly quickly
20:56:56 <Deewiant> fungot: Quite accurate, actually!
20:56:56 <fungot> Deewiant: i don't really care about the finer points of old crotch blended highland scotch whiskey.
20:57:01 <oerjan> fungot: oh come on it's not _that_ weird
20:57:01 <fungot> oerjan: it's very awesome
20:57:28 <oerjan> something psygnisfive would like, i'm sure
20:57:44 <Deewiant> fungot: so wait, you don't care about the /finer points/?
20:57:45 <fungot> Deewiant: if you can turn this into
20:58:10 <fizzie> fungot just wants to get drunk, he doesn't care about the finer points of the crotch-blended whsikey.
20:58:10 <fungot> fizzie: it has the same illness as i do
20:58:26 <Deewiant> fungot: I did not need to know that.
20:58:27 <fungot> Deewiant: but printing ')'
20:58:32 <fizzie> Not everything is all right with that bot.
20:58:45 <oerjan> fizzie: mispleing whiskey is a bad omen
20:59:01 <Deewiant> Drinking alcohol blended with old crotches and then getting illnesses from it? No, everything is certainly not right.
20:59:24 <oerjan> Deewiant: now now, remember strong alcohol is a disinfectant. i think.
20:59:37 <Deewiant> oerjan: not if you get illnesses from it it isn't :-P
21:00:00 <fizzie> fungot: Do you think you'd feel better after a rewrite?
21:00:01 <fungot> fizzie: are/ were any problems with hard disk space on a measly 40 gb drive.
21:00:36 <fizzie> fungot: Actually you only have a ~20 GB drive, but don't worry, only the language models take up much space.
21:00:36 <fungot> fizzie: too slow to perform the o(n) resize once every n inserts
21:01:33 <oerjan> fungot: are you doing something fizzie doesn't know about?
21:01:33 <fungot> oerjan: egobot does not flood.
21:07:08 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | let __ = __ in __ :: t.
21:07:55 <oerjan> ye olde infinite loope
21:08:28 <oerjan> (actually probably trapped)
21:09:02 <oerjan> ghc traps simple infinite loops
21:09:25 <oerjan> those that reevaluate the exact same expression
21:10:35 <fizzie> SWI-Prolog has an amusing easter egg if you ask it a query like "X." but I've probably mentioned it on this very channel already.
21:10:46 <oklopol> i realized the other day it's actually quite trivial to notice you're reevaluating something, after you do it once, you're in a cycle, and it's enough to store one state in the cycle, and check if it appears again
21:10:56 <oklopol> it reappears iff there's a loop
21:11:37 <fizzie> The cycle might start later than that one state you've stored.
21:11:48 <oklopol> yes, but it's enough to change the stored state every now and then
21:11:58 <AnMaster> fizzie, I believe you were saying you would have used ATHR and so on?
21:12:02 <fizzie> Then you might not notice the cycle if it's long enough.
21:12:03 <AnMaster> fizzie, http://rafb.net/p/YpkaJU36.html may interest you
21:12:16 <oklopol> that's actually quite true
21:13:46 <psygnisfive> i think my formalism might be equivalent to CCGs
21:13:50 <fizzie> If I have free time and the inclination to do a fungot rewrite, I might consider some form of ATHR-style threading.
21:13:51 <fungot> fizzie: it's only the html pages are served up using lighttpd, and the
21:14:08 <AnMaster> fizzie, be aware of that efunge is slower than even rc/funge
21:14:29 <AnMaster> it is more for "interest feature ideas" than "raw performance"
21:15:09 <fizzie> Well, the IRC thing isn't really very speed-critical. If something's too slow to implement on the Brainfuck interp, I can do it as a "native" command.
21:15:11 <AnMaster> and cfunge will never have all those weird fingerprints. Just the more tame ones.
21:15:42 <AnMaster> fizzie, also SOCK hm, I will probably do my NSCK idea (which fixed lots of issues with SOCK/SCKE)
21:15:57 <AnMaster> maybe SOCK too, but it was kind of messy to implement
21:18:35 <fizzie> Well, we'll see. I may start simply by cleaning up the existing code a bit. And I still lack the good editor.
21:32:58 <fungot> okokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokokoko ...
21:35:18 <lament> banananananananananonokokokokokokokokokokoko
21:36:42 <oklopol> numbers N for which doing N<o>N for any <o> a hyper operator is the identity function?
21:37:00 <oklopol> not identity, but all produce the same result
21:37:44 <oerjan> what is a hyper operator?
21:38:10 <oerjan> oh you mean the ones i said
21:38:25 <oklopol> 2+2, 2*2, 2^2, 2\/-2, 2&"2, ...
21:39:22 <oklopol> also 0^0 isn't usually defined afaik
21:39:28 <fizzie> Our high-school mathematics teacher used to say 0^0 is mickey mouse with one ear missing.
21:39:46 <fizzie> (Meaning: not very defined.)
21:39:51 <pikhq> 0^^(0^0), on the other hand...
21:39:56 <oerjan> no but i vaguely recall discussions that said 1 is the most reasonable value
21:41:17 <fizzie> A feasibool is like a bool value, but it can only take values that are (semantically speaking) feasible.
21:41:52 <fizzie> According to p. 408 of Knuth (1992), [0^0] "has to be 1".
21:42:17 <fizzie> Well, if Knuth says so, who am I to argue.
21:42:36 <fizzie> It's the "appeal to authority" method of proof.
21:43:07 <ehird> lol - <lilo> [Global Notice] Hi all. At 19:30 UTC, in two hours, we'd like to ask everyone to observe a minute of silence in sympathy with the victims of the terrorist attacks on September 11, their loved ones and friends. Channel admins, if you'd like to participate, please +m your channel for a minute and optionally deop at that time. Thanks.
21:43:19 <fizzie> Or is it "Proof by eminent authority"? The example is given as: "I saw Karp in the elevator and he said it was probably NP-complete."
21:44:11 <ehird> fizzie: proof by knuth
21:44:24 <fizzie> Oh, it has a separate category.
21:45:30 <oerjan> http://en.wikipedia.org/wiki/Karp%27s_21_NP-complete_problems
21:45:59 <fizzie> That page was in my browser cache.
21:46:33 <oerjan> mine too i assume, since i saw it yesterday
21:51:16 <psygnisfive> i found the perfect outlet for my computery languageu urges!
21:59:31 <olsner> computational linguistics? has an IRC channel?
22:00:03 <fizzie> Probably a support group after that mean xkcd strip.
22:01:27 <GregorR> I've also found an outlet for my computer language-y urges.
22:01:31 <GregorR> Programming language research.
22:07:23 <olsner> also known as bantering in #esoteric? :P
22:07:59 <psygnisfive> i mean language-y in the natural language sens,e gregorr :P
22:09:22 <oklopol> olsner: no he's moved on to the real world now
22:09:44 <olsner> oh, real world... how boring :)
22:09:45 <ehird> <psygnisfive> i found the perfect outlet for my computery languageu urges!
22:10:29 <olsner> 'computery' = kindergarten 'computational'
22:10:45 <oklopol> computationalative languagation
22:11:45 <GregorR> Computationalaxative lenguanation
22:14:12 <oerjan> de linguis non est computandum
22:14:45 <oklopol> computationalativatiosivecious
22:16:03 <oklopol> you can put any number of suffices on an english root, and it'll be pretty and cute.
22:16:45 <psygnisfive> i'd put my suffix on your root, if you know what i mean
22:17:34 <oklopol> you have too ciouses and ives too
22:17:57 <oklopol> ARE YOU PLAGIARIZING MY FAILURE
22:18:22 <GregorR> I'll put my suffix on YOUR root.
22:18:40 <oklopol> i'll put my root on your *mother*'s suffix
22:19:13 <oklopol> ...err do gays have mothers, actually?
22:19:32 <GregorR> They sprout from rocks and/or eggs.
22:19:53 <psygnisfive> hence our affinity for metal and metal-related things
22:20:49 <oklopol> GregorR: we all know you gods hate fags
22:21:05 <GregorR> oklopol: Hey I swapped my personality back.
22:21:15 <psygnisfive> unless its on a guy, in which case its powers of rash are reduced significantly
22:22:06 <GregorR> No, it reverts automatically after a timeout.
22:22:11 <oklopol> this is why all the gayness has been bottled up tonight, we were scared of you
22:22:36 <oklopol> i should get back to reading
22:22:39 <oerjan> does that mean we can cancel the protest against you?
22:23:05 <oklopol> it's just the book is excruciatingly hard to read :P
22:23:11 <GregorR> http://codu.org/pics/other/pec2.jpg
22:24:12 <oklopol> psygnisfive: still the same book
22:24:32 <psygnisfive> now maybe add a long shadow-reflect with the same angle as the shadow-reflect of your right leg
22:25:09 <oklopol> algorithm design by eva tardos and jon kleinberg
22:25:32 <oklopol> read the first 600 pages for a course, but need to read the rest for another one
22:25:44 <oklopol> and the last few hundred pages are complete mindfuck
22:25:46 <psygnisfive> im going to read some lecture notes (essentially a book) from an MIT math-for-CS class
22:25:54 <psygnisfive> so its all discrete math and combinatorics and stuff
22:25:59 <oerjan> she never wrote another book, since that would be retarded
22:26:10 <oklopol> (and no one dl this somewhere and tell me it's simple or i will slap you with my trout)
22:26:42 <oklopol> i just have it in book form
22:26:48 <oklopol> and i managed to destroy even that
22:26:55 <oklopol> by soaking it in water for about a day
22:27:14 <oerjan> oklopol: you have a trout?
22:27:20 <oklopol> i get about 200e a month, 78.6 euros for a book, and i destroy it in a week :)
22:27:38 <psygnisfive> yeah but you live with your parents, oklopol
22:27:42 <oklopol> that's how i use the remaining 121.4e
22:28:02 * oklopol is gay and lives with his parents
22:28:20 <psygnisfive> how can you be gay, you have a girlfriend, dont be silly
22:28:31 <oklopol> i doubt i've lived with my parents during the time you've known me
22:28:55 <olsner> you've both been here longer than that haven't you?
22:29:06 <psygnisfive> i dont know if ive been here since february
22:29:13 <olsner> but you *did* live with your parents! hah!
22:29:48 -!- ehird__ has quit ("Leaving").
22:29:48 -!- ehird_ has quit ("Leaving").
22:30:07 <oklopol> but my parents live in this city, i do get moneys from them if i need
22:30:26 <oklopol> i don't, though, 200 is enough for my needs
22:30:32 <psygnisfive> they killed people by forming queues at convenience stores at 3am
22:31:02 <olsner> or a queue being de-queued
22:31:04 <oerjan> yay! now i can sharpen my razors
22:32:23 <oerjan> http://en.wikipedia.org/wiki/Pyramid_power
22:32:25 <psygnisfive> have i mentioned that i despise those people?
22:32:50 <fizzie> Added back that missing echochohoo.
22:33:15 <oerjan> psygnisfive: i thought you liked balls with nuts
22:34:07 * oerjan thought the nuts were the things inside the balls
22:34:26 <fizzie> Put the nuts in the pyramid, you'll get them sharpened.
22:34:40 <olsner> wait, what, no, the nuts *are* the balls
22:34:43 <psygnisfive> http://web.mac.com/arnold_zwicky/BizarroErrors.gif lulz
22:36:02 <olsner> that's a quite severe invasion of privacy there, psygnisfive
22:36:20 <olsner> no knuffeling allowed without permission
22:37:40 <olsner> only -ish? then you can't be exhausted!
22:37:52 <olsner> you'd have to settle for very tired, IMO
23:01:05 -!- olsner has quit ("Leaving").
23:17:43 -!- Corun has quit ("This computer has gone to sleep").
23:18:15 -!- LinuS has quit ("Puzzi. Sì, parlo proprio con te. Puzzi.").