01:18:20 <fizzie> There was a girl scout cookie selling stand at the BART/Caltrain station in Millbrae.

01:19:55 <oren> See my windows installation was totally fine, up until... I installed the newest flash player-- which came with McAfee for some reason.

01:22:52 <oren> My response to thatsort of thing is to boot into linux and DELETE EVERYTHING that isn't system or created by me

03:34:12 <Sketra> Me and my friends are discussing how many human bodies you can fit in a basement and we went from dismembering to boxing to liquifying

05:00:49 <HackEgo> [wiki] [[BrainClub]] http://esolangs.org/w/index.php?diff=42099&oldid=33850 * Zzo38 * (-109) Broken link removed

05:01:13 <HackEgo> [wiki] [[BrainClub]] http://esolangs.org/w/index.php?diff=42100&oldid=42099 * Zzo38 * (-104) Broken link removed

05:56:11 <zzo38> The documentation for Turtle says in section 2.7 that a blankNodePropertyList is allowed as a subject, but section 6.5 says it isn't allowed as a subject; it can only be used as an object.

17:21:32 <mroman> although the documentation will be out of date since all the errors in it people reported I just fixed on the live one

19:35:24 <zzo38> Your chess elo sank below 1000 in what organization of keeping track of these though?

20:13:04 <tswett> Have you ever wanted to know the simplest rational approximations to all the multiples of 1/100 from 0 to 1/2?

20:14:07 <int-e> (I'm slightly surprised that fn*rdbot is still around. It has no clue about reconnecting.)

20:19:33 <fnordbot> boily: got to go. " i don't want to go to work at," and he turned. " but who the man with the five heads all the tests, learning to distinguish between him pretending to be an out a wild whoop in major thirds, threw ford prefect, " it's dark," he was saying, " magrathea is a myth, a fairy story, it's what parents tell their kids at school nicknamed him ix, which in the language. ford was running after him very fast? " very pret

20:20:03 <tswett> I guess I should have "simplest rational numbers that round to all two-digit decimals from 0.00 to 0.50".

20:20:52 <fnordbot> ais523: would work, a mere nothingth of a second, and then his mind to have good ideas with, it must be said, some success. " the babel fish," he was saying, " magrathea is a myth, a fairy story, it's what parents tell their kids at school nicknamed him ix, which in the language. " a computer whose merest operational parameters i am not worthy to calculate the trajectory of every single dust particle of the ship. " it hardly ma

20:21:17 <ais523> also, at least one of the inputs to that style is the hitchhiker's guide to the galaxy

20:35:28 <int-e> right. it's not inconsistent, but when truncating a continued fraction decomposition, you also have to consider replacing the last number by a smaller natural number.

20:38:24 <tswett> > let approx t = do denom <- [1..100]; num <- [0..denom]; guard (round (num/demon) * 100 == t); return (num, demon) in do target <- [0..50]; let (x,y) = head (approx target); show x ++ "/" + show y ++ "; "

20:38:55 <int-e> Err, whatever, 0.19 = 1/(5+1/(3+1/(1+1/4))); 3/16 = 1/(5+1/3), so it is actually a proper truncation. The theory is a bit more complicated than that anyway, because truncating is not always the right thing to do; I think that sometimes one has to increase the last term by one.

20:39:46 <tswett> http://en.wikipedia.org/wiki/Continued_fraction#Best_rational_approximations - decrease it by one.

20:40:14 <tswett> > let approx t = do {denom <- [1..100]; num <- [0..denom]; guard (round (num/denom) * 100 == t); return (num, demon)} in do {target <- [0..50]; let (x,y) = head (approx target); show x ++ "/" + show y ++ "; "}

20:40:31 <tswett> > let approx t = do denom <- [1..100]; num <- [0..denom]; guard (round (num/denom) * 100 == t); return (num, demon) in do target <- [0..50]; let (x,y) = head (approx target); show x ++ "/" + show y ++ "; "

20:42:49 <int-e> tswett: Oh. I was thinking about how to get 1/8 from 0.13 = 1/(7+1/(1+1/(2+1/4))), but of course that's 1/(7 + 1/1).

20:44:23 <tswett> > let {approx t = do {denom <- [1..100]; num <- [0..denom]; guard (round (num/demon) * 100 == t); return (num, demon)}} in do {target <- [0..50]; let (x,y) = head (approx target); show x ++ "/" + show y ++ "; "}

20:45:42 <zzo38> Well, there is also a "inner let" in the do-block, maybe something is wrong wht that

20:46:55 <quintopia> ais523: as the number of states and/or symbols goes to infinity, would you expect the probability that a randomly chosen TM is universal to go to one, zero, or some other constant?

20:47:54 <tswett> > let approx t = do denom <- [1..100]; num <- [0..denom]; guard (round (num/denom) * 100 == t); return (num, demon) in do target <- [0..50]; (x,y) <- return (head (approx target)); show x ++ "/" + show y ++ "; "

20:47:57 <int-e> Koen_: no, but it doesn't surprise me too much. I mean, it's just x^3/(1-x-x^2) with x=1/10... Oh you must have fib(0)=1, who does that?

20:48:02 <tswett> > let approx t = do denom <- [1..100]; num <- [0..denom]; guard (round (num/denom) * 100 == t); return (num, denom) in do target <- [0..50]; (x,y) <- return (head (approx target)); show x ++ "/" + show y ++ "; "

20:48:06 <Koen_> quintopia: you can actually prove that the probability that a string of symbols follows a simple 'logic' is very very very low

20:48:10 <ais523> basically because in order to prove a language with a lot of commands TC, you're normally trying to find a TC subset

20:48:40 <ais523> but it's unclear how likely a randomly generated TM is to have well-behaved subsets (that you can keep within the subset)

20:48:43 <quintopia> but i can't process the decrease rate of the probability of a submachine being "self-contained"

20:49:10 <tswett> > let approx t = do denom <- [1..100]; num <- [0..denom]; guard (round (num/denom) * 100 == t); return (num, denom) in do target <- [0..50]; (x,y) <- return (head (approx target)); show x ++ "/" ++ show y ++ "; "

20:50:08 <quintopia> ais523: i feel like the probability decrease should be roughly on par with the increase in number of submachines though. I lean towards "other constant"

20:50:23 <tswett> > [ let (x,y) = head . filter (\(x,y) -> round (100*x/y) == n) $ rats in show x ++ "/" ++ show y | n <- [0..50] ]

20:50:40 <ais523> I guess a simpler question is "what probability is it that there's a straight simulation of any given TM in a randomly chosen TM?"

20:50:56 <tswett> > let rats = [ (x,y) | y <- [1..100], x <- [0..y] ] in [ let (x,y) = head . filter (\(x,y) -> round (100*x/y) == n) $ rats in show x ++ "/" ++ show y | n <- [0..50] ]

20:50:57 <lambdabot> ["0.0/1.0","1.0/67.0","1.0/40.0","1.0/29.0","1.0/23.0","1.0/19.0","1.0/16.0"...

20:51:42 <int-e> Koen_: and in any case the real point was that this is fairly easy to verify with generating function theory. And the same theory also allows one to construct such "surprising" identities.

20:51:46 <Koen_> int-e: I discovered that when I was 13 and I was quite puzzled. I was even more puzzled when I realized rational numbers have a periodic decimal development

20:52:48 <int-e> Koen_: Oh I can imagine that. Adding lots of random-looking-numbers, why would the result ever be periodic...

20:55:00 <int-e> Koen_: and it probably becomes *more* mysterious if you know the (((1+sqrt(5))/2)^n - ((1-sqrt(5))/2)^n) / sqrt(5) formula :)

21:00:09 <int-e> Koen_: (The nice thing about starting with fib(0) = 0 is that then fib(n) divides fib(m) if (and only if) n divides m.

21:18:08 <int-e> It should work out for n != 2 though. The real identity I was after was gcd(fib(n),fib(m)) = fib(gcd(n,m)), and that is strong enough to get the "only if" direction except for n = +/-2.

22:42:51 <oerjan> <int-e> Koen_: and it probably becomes *more* mysterious if you know the (((1+sqrt(5))/2)^n - ((1-sqrt(5))/2)^n) / sqrt(5) formula :) <-- hm? that immediately gives you that the other one is also the sum of geometric series, so you can calculate it...

23:02:33 <oerjan> boily: are you trying to understand danish numerals that's the only thing relevant that pops into my mind here

23:03:22 <boily> oerjan: me? absolutely not. what would make you think that? and no, I'm not hiding any mapoles behind my back, no sir.

23:04:53 <oerjan> boily: because in danish, 58 = otte og halvtreds, which basically means that formula

23:07:14 <oerjan> Sketra: i'm afraid danish is only fun when you can hear it pronounced, which doesn't work too well over irc

23:11:20 <FireFly> Here's a good documentary on the danish language: https://www.youtube.com/watch?v=s-mOy8VUEBk

23:11:30 <oerjan> why the fuck has youtube started insisting on playing everything as a video series :(

23:16:37 <Sketra> Washington state, Tacoma, WA; South puget sound AVE house number unknown: Weight 123 Lbs.

23:18:09 <boily> a precise and timely answer??? what has this chännel come to... oh the fungotity...

23:19:03 <zzo38> Shouldn't you indicate using degrees of longitude/latitude? Possibly you might also want the amount of elevation

23:20:27 -!- Koen_ has quit (Quit: The struct held his beloved integer in his strong, protecting arms, his eyes like sapphire orbs staring into her own. "W-will you... Will you union me?").

23:20:39 <Phantom_Hoover> even then you can get the information from surface elevation at a given point

23:20:44 <zzo38> I thought you were trying to calculate the center of gravity or something like that, in such case you do require elevation too

23:21:04 <zzo38> And the elevation from coordinates might not be good enough since it might be a tower

23:21:08 <FireFly> Or if you want to take the y-position of people living in skyscrapers into account

23:21:47 <zzo38> You don't need the elevation if you are trying to plot a horoscope, but then you don't need your weight either, and anyways that doesn't seem what you try to make

23:22:32 <FireFly> boily: do you have the list of approximate coördinates and body weight somewhere?

23:22:42 <Sketra> window had a cat named Peru who watched the elevation of earth grow steadily sideways?!?!!? We like water melons and eat grains of universes

23:23:31 <zzo38> oerjan: You already can use the sun as center of measurement if you want to; it doesn't have to be Earth at the center

23:23:40 <oerjan> i wonder what kind of mess the astrologers will make once we do start colonizing other planets

23:23:58 <zzo38> However, no software I have seen can calculate houses when the center of meausurement isn't the Earth.

23:24:48 <zzo38> Nevertheless, if you are able to define a zero longitude for the planet or other object, you should be allowed to calculate houses too.

23:26:41 <zzo38> oerjan: Using a spaceship as the center of measurement will be more difficult; there doesn't seem to be any way to reasonably do that.

23:27:48 <oerjan> zzo38: my idea is that you "really" have a house-like system for each planet, not just the one you're on, it's just that it's irrelevant when the distance to it is huge

23:34:29 <oerjan> boily: i also have reason to believe that Sketra and Lilax are very very close hth

23:36:22 <Sketra> Excuse me oerjan but I'm much more structured than Lilax although you aren't wrong about the body I'm just more neat and well organised rather than lilax where as he likes to fuck with everything and annoy oerjan

23:38:14 <Jafet> Islam seems to have worked something out http://archive.wired.com/science/space/news/2007/09/mecca_in_orbit

23:39:49 <oerjan> well you take the slices and then you put the smallest one on the left and the biggest one on the right

23:43:34 <oerjan> boily: i seem to be unable to follow the discussion without jumping through more hoops than i have patience for.

23:59:28 <pikhq> Yeah. The typical Japanese "r" is a tapped "r", and it's a quick step from there to rolling it.