1. Joined
    08 Oct '06
    Moves
    24000
    01 Apr '10 06:441 edit
    Any rational number can be expressed as a turnacating decimal, or a non-turnacating decimal which eventually turns in to an infinitely repeating sequence of numbers. Does anyone know why that last fact is true? Just something I'm curious to know.

    Furthermore, any decimal number which ends with an infinitely repeating sequence of numbers is rational. I would also like to know why this is.

    Also, is this a peculiarity of the decimal number system, or is this true in any base?
  2. Joined
    07 Sep '05
    Moves
    35068
    01 Apr '10 07:36
    The second part is fairly straightforward. Here's a demonstration, but the principle would apply to any such number.

    Let's say we've got the number 0.abcdefdefdef...(def repeats infinitely)

    = 0.abc + def x 10^-6 + def x 10^-9 + def x 10^-12...

    = 0.abc + def x 10^-6 x (1 + 10-3 + 10-6 + ...)

    But that series in the brackets is a geometric series. We know what that adds up to:

    1 + 10^-3 + 10^-6 + ... = 1/(1 - 10^-3), which is rational (= 1000/999)

    So we've now got an expression that's just a (finite) sum of things we know is rational. Which must be rational. Not only do we know any such number is rational, it's not all that hard to work out what it is as a fraction.

    There's also nothing in there that would work differently in other number bases - your geometric series would be different, that's all.
  3. Joined
    07 Sep '05
    Moves
    35068
    01 Apr '10 09:23
    Incidentally: "turnacating"? I can see what you mean from context, but I've never heard that word before. It's not in the OED, and Googling doesn't have anything relevant. You sure that's right?
  4. Joined
    07 Sep '05
    Moves
    35068
    01 Apr '10 09:35
    For the first part, think about how you calculate the decimal expansion of a fraction. Let's say:

    X/Y = 0.abcdefghijklm....... (where X & Y integers, and X < Y)

    We can obtain each decimal by multiplying by 10 and taking the integer part. Then subtract that:

    (10X - aY)/Y = 0.bcdefghijklm........

    and repeat.

    Each time, the left-hand side is a fraction < 1 with Y as the denominator. But either:

    1. It hits zero. In which case we stop, and we've got a finite-length ('turnacating'?) decimal, or

    2. Since there are only Y-1 possible such non-zero fractions, eventually we have to hit one we've had before. And the pattern repeats from this point on.
  5. Germany
    Joined
    27 Oct '08
    Moves
    3081
    01 Apr '10 09:51
    You can also use that nifty trick to show that 0.999... = 1.
  6. Standard memberPalynka
    Upward Spiral
    Halfway
    Joined
    02 Aug '04
    Moves
    8702
    01 Apr '10 10:16
    Originally posted by mtthw
    Incidentally: "turnacating"? I can see what you mean from context, but I've never heard that word before. It's not in the OED, and Googling doesn't have anything relevant. You sure that's right?
    I also never heard the expression, I think in English the correct expression is recurring decimal.

    Good work with the other posts.
  7. Joined
    08 Oct '06
    Moves
    24000
    01 Apr '10 13:09
    Wow, I made up a word. That's new. I think the word I was looking for was truncate, or something. Thanks for the posts.
  8. Germany
    Joined
    27 Oct '08
    Moves
    3081
    01 Apr '10 20:30
    I think you meant "non-truncated".
  9. Joined
    08 Oct '06
    Moves
    24000
    01 Apr '10 22:331 edit
    Yea.

    Can I use my upcoming quantum test as an excuse for this rather glaring mistake?
Back to Top