Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
×
Math Graphics Science

Mandelbrot Zooms Now Surpass the Scale of the Observable Universe 157

StartsWithABang writes You're used to real numbers: that is, numbers that can be expressed as a decimal, even if it's an arbitrarily long, non-repeating decimal. There are also complex numbers, which are numbers that have a real part and also an imaginary part. The imaginary part is just like the real part, but is also multiplied by i, or the square root of -1. It's a simple definition: the Mandelbrot set consists of every possible complex number, n, where the sequence n, n^2 + n, (n^2 + n)^2 + n, etc.—where each new term is the prior term, squared, plus n—does not go to either positive or negative infinity. The scale of zoom visualizations now goes well past the limits of the observable Universe, with no signs of loss of complexity at all.
This discussion has been archived. No new comments can be posted.

Mandelbrot Zooms Now Surpass the Scale of the Observable Universe

Comments Filter:
  • Ehhh What ? (Score:5, Insightful)

    by Crashmarik ( 635988 ) on Sunday April 19, 2015 @04:28PM (#49506455)

    Technically the description of the Mandlebrot set is encoded within the observable universe so there is a problem in recursion her.

    Second how is this surprising to anyone ? It's long been possible to describe and mathematically manipulate sets with more elements than the observable universe.

    • Re:Ehhh What ? (Score:5, Informative)

      by disputationist ( 1324927 ) on Sunday April 19, 2015 @04:43PM (#49506525)
      Incorrect. Abstract mathematical objects are not "encoded within the observable universe"
      • Re:Ehhh What ? (Score:4, Interesting)

        by ultranova ( 717540 ) on Sunday April 19, 2015 @09:47PM (#49507905)

        Incorrect. Abstract mathematical objects are not "encoded within the observable universe"

        Sure they are. The set of concepts that humans can conceive are those which human brains, either directly or through tools like computers, can handle. Human brains evolved in the context usually called "the observable universe", so all concepts - including but not limited to abstract mathematical objects - we can think about are encoded within it, just in a real roundabout way. In other words, you can not know anything that isn't encoded in your causal past; even the very notion of abstraction only exists because it's inherent in the physical universe to such a degree that evolution encoded the principle into your brain.

        And besides, the notion that math is supernatural - something that exists above physical reality, independent of it - is an unproven and probably unprovable assertion.

        • by gl4ss ( 559668 )

          math, as a set of rules and logical conclusions made from them, doesn't depend on the universe. that's whats magical about it. some alien force should come to same math conclusions, including mandelbrot set.

          it's not "above" physical reality, it's more like parallel.

          it's a real shame that the voyager doesn't include a mandelbrot set.

          • math, as a set of rules and logical conclusions made from them, doesn't depend on the universe.

            I'm not sure that's provable, as GP said, especially when the term "magical" is invoked as a descriptor. It quickly becomes a philosophical argument, and without a testable hypothesis, probably not worth debating.

            • Math doesn't depend on the Universe. It happens that some mathematical constructs are extremely useful in modeling the Universe, and over the centuries we've tended to concentrate on the more practically useful varieties of math. There's nothing magical about it. Math isn't physics and physics isn't math, but they get very intertwined sometimes.

              • > Math doesn't depend on the Universe

                are you saying that if the universe didn't exist then Math could still exist?

                that's a bold statement, and i'm not sure you have a proof.

                • Math depends on logic. Without some sort of Universe, we wouldn't have anything that can do logic. That's the limit of the dependence, since math has nothing to do with the nature of the Universe.

                • by Xest ( 935314 )

                  It's not an unreasonable viewpoint given that we can use math to describe universes that physically could not exist.

                  Math obviously exists outside of those particular universes, thus, one must reasonably conclude that either math can exist outside of any particular universe, or that for some reason some universes, such as ours (or perhaps only ours), are special cases where math exists.

          • by Alsee ( 515537 )

            it's a real shame that the voyager doesn't include a mandelbrot set.

            They were going to include one, but they were unable to complete it by the launch date.

            -

        • you're confusing conceptual and abstract. they're different.

          concepts are the components of thought, and require a mind.

          logic (math) is abstract and does not require a mind.

          the question of whether or not logical absolutes can exist without a universe is probably not a useful one to consider.

        • Incorrect. Abstract mathematical objects are not "encoded within the observable universe"

          Sure they are. The set of concepts that humans can conceive are those which human brains, either directly or through tools like computers, can handle. Human brains evolved in the context usually called "the observable universe", so all concepts - including but not limited to abstract mathematical objects - we can think about are encoded within it, just in a real roundabout way. In other words, you can not know anything that isn't encoded in your causal past; even the very notion of abstraction only exists because it's inherent in the physical universe to such a degree that evolution encoded the principle into your brain.

          And besides, the notion that math is supernatural - something that exists above physical reality, independent of it - is an unproven and probably unprovable assertion.

          You are confusing mathematics with the metalanguage we use to describe them.

    • Re:Ehhh What ? (Score:5, Informative)

      by X0563511 ( 793323 ) on Sunday April 19, 2015 @04:47PM (#49506555) Homepage Journal

      The set is not encoded in the universe, though the description of the set is. Else, every reference to "infinite" would, well, break the universe.

      • by JustOK ( 667959 ) on Sunday April 19, 2015 @04:55PM (#49506593) Journal
        And you think the universe isn't broken NOW? Good god, man. Wake up!
      • LOL do you run 5 accounts to mod yourself ?

        The set is encoded in a one line, that defines every point in it.

        • by Pav ( 4298 )
          Unicorns exist... I've encoded them in my mind.
        • by gl4ss ( 559668 )

          but the result isn't known until you calculate it.

          you could use all the energy in the world to calculate it and still not finish calculating the set. that's the how every reference to infinite would break the universe.. the definition is in this universe, BUT the results are not calculated unless someone calculates them.

          comparing the result to complexity of the universe is a bit silly though since mandelbrot as a set you could zoom infinitely AND _never_ find an image of the universe or billy gatesy(though

          • The set is complex not random. You'd no more expect particular random images in it than you would an indefinitely iterated sierpinski gasket.

            Anyway at this point I am guessing is that "Ask Ethan" made friends with somebody at Slashdot. Which explains why these non news non stories keep showing up here.

            • Depends on your definition of "random", I guess. Under certain circumstances, it's unpredictable without calculating it, but it can be calculated.

              • Well It is a connected set so that rules out an infinite number of images without bothering to actually calculate or look for them.

    • Second how is this surprising to anyone ? It's long been possible to describe and mathematically manipulate sets with more elements than the observable universe.

      That's what I was wondering. Even going to the extreme, the diameter of the universe is about 5x10^61 Planck lengths. This is the sort of figure mathematicians have been happy to play for years now.

  • by ma++i+ude ( 580592 ) on Sunday April 19, 2015 @04:30PM (#49506461) Homepage
    A zoom into a fractal stored as a 16-minute YouTube video must be the least efficient way to store an equation. If only there was some sort of a 'fractal compression' method.
    • Re:YouTube? Srsly? (Score:5, Insightful)

      by itzly ( 3699663 ) on Sunday April 19, 2015 @04:36PM (#49506491)

      If only there was some sort of a 'fractal compression' method.

      I'm looking forward to your decompressing code that can reproduce the video in less than 16 minutes.

    • Re: (Score:2, Funny)

      by Anonymous Coward

      Hi. Welcome to CS 121. Today we discuss the Time vs Space complexity tradeoff.

    • A zoom into a fractal stored as a 16-minute YouTube video must be the least efficient way to store an equation. If only there was some sort of a 'fractal compression' method.

      Plus, the article states that they only zoom in by a Google squared... presumably because Google set that limit for YouTube.

  • Can we peak under alien skirts, or is this only virtual?

  • by Hognoxious ( 631665 ) on Sunday April 19, 2015 @04:49PM (#49506563) Homepage Journal

    It's not n^2 + n, it's n^2 + c.

    That's to say, the number you multiply by itself isn't the same as the number you add.

    • It's not n^2 + n

      Yes it is, for the second (or third, if you're starting from 0) element in the sequence. The article isn't defining the sequence, per se; it's listing elements in the sequence calculated solely from the initial complex number.

      I think the confusion has arisen because n is usually used as the element number, not the complex point (which usually goes by c).

      the number you multiply by itself isn't the same as the number you add.

      No - well, just once - but that's not what the article says. You square the previous element, then add c.

      Wikipedia says:

      The Mandelbrot set is the set of complex numbers 'c' for which the sequence ( c, c^2 + c, (c^2+c)^2 + c, ((c^2+c)^2+c)^2 + c, (((c^2+c)^2+c)^2+c)^2 + c, ...) does not approach infinity.

      which is exactly what the article sa

      • by Mateorabi ( 108522 ) on Sunday April 19, 2015 @05:36PM (#49506805) Homepage
        Some of the confusion is that the original description is defined recursively in a way that 'c' only shows up once, and the initial value is not c. z[i] = z[i-1]^2+c. But because z[0] is defined = 0, you can effectively rewrite the sequence in terms of just 'c' starting from the second. The downside is that at first it might LOOK at first glance like the previous term is being added, which is why I like the recursive form.

        Also, by not starting from 0 you miss out on a cool connection: for a given fixed C, the graph of convergence for non-zero choices of z[0] over the complex plane gives you a Julia Set. With the neat property that Julia Sets from C inside the Mandelbrot set are fully connected and Julia Sets from C outside the Mandelbrot Set are sparse disconnected Cantor spaces.
      • The article isn't defining the sequence, per se; it's listing elements in the sequence calculated solely from the initial complex number.

        You can't do it from one complex number.

        If what you said was true then why does every implementation - and I've written at least two[1] - use two complex variables? And why is there such a thing as a Julia set, the difference being whether it's n (should be z anyway) or c that represents the point on the Argand diagram you're going to colour?

        http://www.fractaldesign.n [fractaldesign.net]

        • You can't do it from one complex number.

          I'm not quite sure what you mean. The summary says (as does Wikipedia) that the sequence goes:

          0
          c
          c^2+c
          (c^2+c)^2+c

          You only need one complex input variable (the coordinate of the point) to determine whether or not the point is in the set. I think your link says as much:

          The calculation of a Mandelbrot set is similar. The difference is in the values that are substituted into the equation. In the equation for f(z) the pixel coordinate (x,y) is substituted into the complex number C and (0,0) is substituted for a starting value of z.

          You can instead just use skip the first iteration and use c as the starting value of z, because that's always the next result after z0=(0,0).

          If what you said was true then why does every implementation - and I've written at least two[1] - use two complex variables?

          Err, I dunno. You wrote them so I'm not sure why you're asking me. If you're just talking about program

  • Old, old news (Score:5, Interesting)

    by wonkey_monkey ( 2592601 ) on Sunday April 19, 2015 @05:02PM (#49506627) Homepage

    Mandelbrot Zooms Now Surpass the Scale of the Observable Universe

    First off, does that even mean anything? What units is the "scale" of a universe expressed in?

    Okay, let's take it to mean the ratio of the size of observable universe to the size of the Planck length, for lack of any better definition. In that case, Mandelzooms surpassed that years ago.

    with no signs of loss of complexity at all.

    You make it sound like we're expecting a loss of complexity, and we just haven't found it yet. But isn't it mathematically proven that the Mandelbrot set has the same "complexity" at all scales? Kind of inherent in the whole "fractal" thing, I thought...

    I'd have thought it would be more interesting to talk about, for example, how all the pretty colours that everyone gawps at aren't even points in the set. They're just colour-coded as to how long the sequence takes to reach a certain value (all of the coloured points ultimately diverge to infinity, which is what makes them not part of the set).

    • First off, does that even mean anything? What units is the "scale" of a universe expressed in?

      I'm a bit rusty in my maths - but I'm fairly certain mega-volkswagons are the currently used scale

    • First off, does that even mean anything? What units is the "scale" of a universe expressed in?

      Scale doesn't have units - if I have a 200x zoom it could be meters or feet or idiotic statements. If only there were an article to answer your fucking questions:

      the largest observable scales are âoeonlyâ 92 billion light years or so (from one edge of the observable Universe to the other), while the smallest theoretical scale, the Planck scale, is down at around 10^-35 meters. All told, this is just 62

      • You're absolutely correct about what the article is asserting, and the GP seems to have overlooked how the scale was determined. At the same time, he did hit upon how the Planck length is an arbitrary divisor.

        "There is currently no proven physical significance of the Planck length; it is, however, a topic of theoretical research."
        https://en.wikipedia.org/wiki/... [wikipedia.org]

        There is no scale to the universe that we can prove.. any length could be infinitely divisible, so it's overselling it a bit to say that the scale

    • Comment removed based on user account deletion
      • Thankyou! I watched the video zooming in and felt that the OP made sense, but didn't know exactly what they meant.

        The only problem was that the OP was too vague, and omitted the numbers - rather than that the concept was stupid.
  • V'ger? Is that you?

  • There are also complex numbers, which are numbers that have a real part and also an imaginary part.

    The movie and recording industries use those for accounting purposes.

  • by PacoSuarez ( 530275 ) on Sunday April 19, 2015 @06:02PM (#49506967)

    For most complex numbers the sequence will most certainly not converge to positive or negative infinity, whatever those mean. When dealing with complex numbers it only makes sense to talk about a single infinity, which is the point at infinity of the projective complex line (a.k.a. "Riemann sphere").

    • This is basically what I came to say. This summary is one of the worst ever.

      Really, one should be talking about approaching infinite absolute value, i.e., distance from the origin (which cannot be negative).

  • by 140Mandak262Jamuna ( 970587 ) on Sunday April 19, 2015 @06:17PM (#49507029) Journal
    Ages ago, seems like bronze age to me now, I was a freshman in college and got my first calculator. A tiny Casio-Fx48 creditcard sized one. It was only 9 decimal digits accurate, but its floating point number range went all the way up to a googol, 9.9999999e+99. That number is so huge, it is more than the number of subatomic particles in the known universe. Ming bogglingly huge number. In math such things are so common. For example the function factorial, reaches a googol at 79. Yup, Factorial (79) > number of subatomic particles in the known universe.

    I read the book "Fun With Numbers" by Mir publications, Moscow in 10th grade. It talked about simple things like immensity of a number like pow(2,64) explained in a simple language a 10th grader could get. (pow(2,64) rice grains would need a barn 3 meter wide, 3 meters tall and several times the distance of Earth to Moon or something like that).

    So Mandelbrot set could exceed the resolution of the known universe, by some version of the definition of these terms, in as little as 64 iterations.

    • by jpatters ( 883 )

      If you think a Googol is big (or a Googolplex), try wrapping your head around Graham's number.

      I'll use "^" to represent a Knuth arrow.
      Start with 3^^^^3, call that g_1.
      Now g_2 is 3^^^...^^^3 but with g_1 Knuth arrows.
      g_3 is 3^^^....^^^3 but with g_2 Knuth arrows.
      G, or Graham's number, is g_64.

      There are numbers with more digits than the number of sub-atomic particles in the universe, that if you repeatedly take the factorial of, over and over again more times than the number of sub-atomic particles in the uni

    • Floating point numbers, by definition, trade accuracy for size. They're compressed as two numbers -- a base and an exponent, and are limited to the precision of the size of the register.

      You need fixed-point numbers to do a level of zoom without losing accuracy, where the level is dependent upon the size of the number you can store.

  • by Overzeetop ( 214511 ) on Sunday April 19, 2015 @06:46PM (#49507139) Journal

    Two hours and nobody has posted this until now: https://www.youtube.com/watch?... [youtube.com]

    It's like you all aren't even trying anymore.

  • https://www.youtube.com/watch?... [youtube.com]

    Relevant part is at 3 minutes and 9 seconds.

    Yes I know, it's "fake", not done in real-time. But it's still an impressive image sequence compression playback that was done on computers 22 years ago*.

    * That was more than two decades ago? Holy shit, I'm old. And get off my lawn!

  • then there is the 3d sets
    like
    http://www.imagebam.com/image/... [imagebam.com]
    http://www.imagebam.com/image/... [imagebam.com]
    or one of my picassa albums
    https://plus.google.com/u/0/ph... [google.com]

    • by GuB-42 ( 2483988 )

      There is no 3d Mandelbrot set.
      What you have are :
      - 4d set using quaternions that is projected to 3d
      - Mandelbulb-like fractals

      Mandelbulb is an extrapolation of the Mandelbrot formula that is tuned to produce pretty pictures, same with other fractals like Mandelbox.
      The quaternion-based set is mathematically closer to the original definition but the pictures it generates are less interesting.

  • That is did he have to zoom in on a very specific point to have content the entire video?

    • by itzly ( 3699663 )

      The Mandelbrot set itself is the collection of points that are shown as black. The set itself is a fully connected, but very complicated, shape. If you zoom in on a point inside of it, after a while you only see black. If you zoom in on a point outside of it, it will become another solid color. In order to keep it interesting, you need to zoom in right on the edge. But the edge is infinitely long, so there are many interesting points where you can zoom in.

  • with no signs of loss of complexity at all

    I should hope not, given that its self similarity and the fractal dimension of its boundary are established mathematical results.

  • by edittard ( 805475 ) on Monday April 20, 2015 @01:48AM (#49508647)
    How can something which is just a pure number outscale something that's physical and has actual dimensions?
    • Re:confused (Score:4, Informative)

      by wonkey_monkey ( 2592601 ) on Monday April 20, 2015 @03:55AM (#49508961) Homepage

      The idea is that the "scale" of the observable universe is the ratio from the largest "thing" (the whole observable universe) to the smallest "thing," which is the Planck length. That ratio is 10^63 or something like that, much less than the zoom level that's achieved in the video.

  • by jandersen ( 462034 ) on Monday April 20, 2015 @03:51AM (#49508945)

    The scale of zoom visualizations now goes well past the limits of the observable Universe, with no signs of loss of complexity at all.

    I have deperately tried to interpret some insight into this 'discovery' - and failed; this may be because of my lack of understanding, of course, but I don't think so. Mathematically, the set of complex numbers is infinite - uncountably so, in fact (Cantor's diagonal argument):

    http://en.wikipedia.org/wiki/C... [wikipedia.org]

    The observable universe is limited by the speed of light, so it will be less than ~28 ly across (we can at most see as far as light has traveled since the big bang), and intuitively infinite must be bigger than something of limited size. It is a misleading argument, though; infinity is a strange thing, and comparing the sizes of infinite sets has to be done with care (as Cantor's argument demonstrates). For one thing, we don't really know that the universe is a continuum in any of the senses defined in mathematics - there are speculations that there is a "smallest size" of distance and time "because of quantum" (I'm being deliberately wooly-mouthed because I don't know what I'm talking about here). If that is the case, then any infinite set will have more elements than there are bits of universe that we can observe (total volume of observable universe / volume of.the smallest element = finite number)

    If we are talking about continua, on the other hand, then we don't really know, I think. A Mandelbrot set is a subset of the complex numbers, so is at most of the same cardinality as that one. Incidentally and perhaps surprisingly, there are exactly as many complex numbers as there are real numbers, and there are as many real number between 0 and 1 as there are between +/- infinity, courtesy Cantor again. The universe, on the other hand may or may not be fully describable as some sort of N-dimensional, smooth manifold (manifold: a winkly version of space, so to speak); a smooth manifold will again have the same cardinality as [0,1], and if the universe can not be fitted into one of those, it is anybody's guess, I think. There are sets larger than the real numbers.

    As an aside note: why have I ignored the idea of 'size' as in distances or volumes? Because it makes no sense to talk about metrics, when one of the sets does not have a defined method of measuring distances in meters or any other physical distance. Assigning a physical unit to an abstract set would be arbitrary.

  • Mandlebrot magnificiation blurs out when you use single-precision floating point. Double precession gets you about another 25 powers of two. I'd go for 128-bit precision to really explore Mandelbrot. Its rarely implemented in hardware or software. http://en.wikipedia.org/wiki/Q... [wikipedia.org]
    • There are many options for arbitrary precision floating point computations. You can easily go to 1024-bit precision (or more) with some easy to find classes for C++ or Java (or write your own). For performance purposes, it's slightly better to go with fixed point arithmetic (especially when your data has known boundaries), but you can get quite reasonable performance from floating point too.
  • Just in for slashdot and in exclusivity, here's a zoom of the Cantor set at 2^1048576 :

    _ _

"The most important thing in a man is not what he knows, but what he is." -- Narciso Yepes

Working...