Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Am I wrong? Is this correct thinking? Am I weird? Am I crazy?

  1. Nov 28, 2011 #1
    http://www.teacherschoice.com.au/maths_library/area and sa/area_2.htm

    "NOTE: The value of p can never be known exactly, so surface areas of spheres cannot be calculated exactly."

    p = pi here.

    It always irks when people say "(insert irrational constant here) doesn't have an exact value" or even "we don't know pi's exact value."

    I don't understand. I can't express it completely as a decimal, so what?

    I know the exact value of pi through many other ways. Series, the square root of gamma function evaluated at 1/2, and at the risk of being yelled at for circular reasoning, the ratio of a perfect circle's circumference to diameter, are all pi. I can only approximate it in decimal form, but saying that means "we don't know it's exact value" or even worse that it "has no exact value" makes no more sense than saying I don't know exactly what my dog looks like because I can't draw an exact picture of him on an etchasketch.

    A simpler way to show the exact value of pi is as follows:


    I mean, I just don't understand this. It's a number, it has a place on the number line, it is clearly defined.

    e is exactly 1/1! + 1/2! + 1/3! ..... so yes, I do know it's exact value. It is a clearly defined number!
  2. jcsd
  3. Nov 28, 2011 #2
    You are absolutely correct. We know pi, e, etc exactly. We cannot represent them nicely as a terminating or repeating decimals in any rational base, but that doesn't change the fact that pi is an exact number.
  4. Nov 28, 2011 #3


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Besides, what if [itex] r=1/\sqrt{\pi} [/itex]
  5. Nov 29, 2011 #4
    1MileCrash, you think like a mathematician :approve:

    Here's a little joke on the topic:
    A physicist, an engineer and a mathematician are asked what pi is. The physicist says that it is [itex]3.14 \pm 1[/itex]. The engineer says that it is approximately 3.1415. The mathematician thinks long and hard. After several hours he proudly exclaims: it is equal to pi!!

    Sorry, your post reminded me of this.
  6. Nov 30, 2011 #5
    So I can draw a perfect circle and measure the length of the line? Neato!
    Last edited: Nov 30, 2011
  7. Nov 30, 2011 #6


    User Avatar
    Gold Member

    But, 1MileCrash:
    But it does change the fact that "we can't calculate the surface of a sphere" with a given radius -which is the reference you mentioned in your OP.

    You'd be able to express it algebraically, but could not give it a unit.

    The measurement of the surface of your sphere will either be 4*pi*r^2 or it will be some value whose error is proportional the square of our imprecision in knowing pi.
  8. Nov 30, 2011 #7


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    You cannot measure circumference or radius to a greater precision then our knowledge of the digits of Pi. So you can indeed calculate the circumference or area. The precision of the result will be determined by your ability to measure, not the value of Pi.
  9. Nov 30, 2011 #8


    User Avatar
    Gold Member

    If I define a hypothetical sphere as precisely 1 inch in diameter, it's circumference and other measurements will only be as accurate as our measurement of pi.
  10. Nov 30, 2011 #9
    That good to see. in my first post I deleted having noticed by looking at my geometry set's circular compass that units defined the "length", so maybe say dividing circumference by diameter does something wonky with the units.
    Last edited: Nov 30, 2011
  11. Nov 30, 2011 #10


    User Avatar
    Gold Member

    I am not really sure what you are trying to say.

    To divide circumference by diameter you must take two measurements. Both will be only as accurate as your measuring device.

    But even if we use ideal, hypothetical objects, wherein we define the measurement as an exact unit, there is still an element of error in what value we use for pi.
  12. Nov 30, 2011 #11
    You're cheating if you say "the sphere is exactly 1 inch diameter" and then in the very next breath limit pi by our measurement accuracy. I don't see how saying something is "exactly 1 inch" and "exactly pi" are any different. "Exactly one inch" and "exactly pi" require the same exact measurement accuracy... infinite.

    Furthermore, if your hypothetical sphere is exactly one inch in diameter not by measurement but by definition, then of course the circumference is exactly pi by definition as well, not an approximation!

    Why couldn't I give a unit? 4pi inches. There is absolutely nothing wrong with that. 4pi is a number. Just like 5, 6, or 2.

    Expressing numbers algebraically is just as valid as writing them out, we express numbers in the most convenient way. I would choose to express a large number algebraically, 6.9 x 10^20 for example, and I could have that many inches. 6.9x10^10 inches, 4pi inches, 8 inches, I don't understand what you're saying.

    If we define the radius to be 1 cm exactly, then the surface area of the sphere is exactly 4pi cm^2 by definition.

    If we measure the radius to be 1 to k amount of significant figures, there is nothing stopping you from measuring just as many significant figures of pi. 1.000 meters means "we stopped measuring after millimeters" and 3.141 meters means "we stopped measuring after millimeters." Exactly 1 meters has just as many decimal places as exactly pi meters.

    Therefore that calculation based on measurement works exactly the same as calculating the area of a rectangle or any other calculation based on measurement that didn't necessarily involve any irrational constant. Take two measurements, and obey significant figure rules.

    If you allow yourself to say "exactly one by definition" then you can say "exactly pi by definition" without a second thought.

    Does that make sense?
    Last edited: Nov 30, 2011
  13. Dec 1, 2011 #12
    Ah ha, either do I really.

    I guess Im thinking of it from the perspective of mutiply two lines of equal length at a right angle make a square. The two dimensional area of the square is calculated exactly. This can't be done with a circle. Said differently units in the x direction are equal to units in the y direction when calculating the area of a square.
    Last edited: Dec 1, 2011
  14. Dec 1, 2011 #13
    Dave is completely correct. You must always consider the element of error with Pi.
  15. Dec 1, 2011 #14


    User Avatar
    Gold Member

    [strike]Units can be exact. One foot is precisely 12 inches to an infinite number of decimal places.

    A square that is one inch on a side is precisely one square inch in area. Zero error.

    What precisely, in square inches, is the area of a circle whose diameter is precisely one inch?[/strike]

    [strike]Yes. But it is unitless. You have only expressed it algebraically.[/strike]

    Hm. Yes it does.

    I was thinking of pi as an algebraic value. It's not.

    "pi" is no different from "one third". They represent precise points on the number line, even if they can't easily be located with a straight pin on that line or be succinctly be represented in decimal form. No one would claim "one third" (or even "1/3") is imprecise or "can never be known exactly".

    Well reasoned 1MileCrash, well reasoned. You are correct and point is conceded.
  16. Dec 1, 2011 #15
    I've learned a lot from your posts, happy to finally return the favor.
  17. Dec 1, 2011 #16
    Irrational constant means you can't represent it using a fraction of real numbers. An infinite series isn't a ratio of a/b (where a and b are real positive numbers, and b != 0), and the gamma function is just an integral which converges absolutely for x>=1*, so that means we can switch the summation and integration signs, which then leads us back to an infinite series, which once again ISN'T a ratio of a/b!

    So yes, you're wrong because you're not sticking true to the definition of an irrational constant.

    *Assuming we're talking about: [tex]{\Gamma (x) = \int^\infty_0 \frac {t^n}{te^t} dt}[/tex]
  18. Dec 1, 2011 #17
    I've never claimed that anything was a ratio of a/b. The true definition of an irrational number is that is *can't* be expressed as a/b where and and b are integers. When have I ever implied otherwise?
  19. Dec 1, 2011 #18


    User Avatar
    Science Advisor

    Nonsense. Every irrational number can be written as a ratio of two reals, namely itself and 1.

    On top of that, the fact that pi doesn't have a periodic decimal representation has no bearing about whether there is a finite representation of pi. Of course it does have one: π. Or you can write everything in base-π. Our choice of decimals is arbitrary anyway.

    MrsLorinda, any error in calculating pi is always dwarfed by the error in your apparatus, for the simple reason we can increase the precision of pi without bound. And if you are not discussing a physical problem, just a hypothetical, then the error of π is always zero. You can argue that that was what you meant in your post, but then you are contradicting Dave (who already retracted his argument).

    Edit: And another thing. One thing that annoys me about the Teacher's Choice article is that they take the square root 39.81 and don't discuss error of that operation. But they needed to point out that pi can "never be known exactly". Huh.
    Last edited: Dec 1, 2011
  20. Dec 1, 2011 #19


    User Avatar
    Science Advisor

    what most people mean by "cannot be determined exactly" (whether they know it or not) is simply that a number is irrational.

    one can view the real numbers as a "perfect measurement system". we, on the other hand, are not-so-perfect. there is a recursion problem with measurement, which is: what do you take as a standard? more precision in measurement requires a very small unit of measurement, and we can only reliably go so small. so in this sense, rational numbers represent an optimal compromise: we can get "close enough" without needing to be "exact".

    so even though we can produce, at will (in theory, our computers ARE limited in power), rational numbers to any desired (finite) degree of accuracy near pi, the "exact" value eludes us.

    the teacher could be seen as wrong, or right, depending on your point of view. even though there is a 1-1 correspondence between positive reals and their magnitudes, the real numbers are just points, sitting there on the number line, whether we know about them or not, but their magnitudes are something we measure, which requires choosing a "unit", which limits us to rational approximations, which are always "precisely wrong" to any desired degree of inaccuracy.
  21. Dec 2, 2011 #20
    You can't draw a perfect circle, but that's a physical limitation. Pi is still a number whose exact value is known. It can be represented in many different ways. It's not a rational number, so we can't represent it as a finite or repeating decimal number in a rational base. Still, pi is exactly...pi or [insert any one of the many series for pi]. I don't understand why we can only "calculate" rational numbers. Surely we know what sqrt(2) is. It is the number whose square is 2. What does it mean to "exactly calculate" something, anyway? The area of a circle is exactly calculated to be pi*r^2. Whether someone wants to round it off and express a number within some arbitrary distance to this number in decimal form is purely up to the person, but we had an exact calculated value before. The teacher's explanation on that site is technically wrong and will probably confuse some of his/her more careful students.

    Right, but even if the teacher meant by "cannot be calculated exactly" as "cannot be expressed as a rational number," that teacher is again wrong because, as office shredder said earlier, if r= 1/sqrt(pi), then the surface area is exactly 4, an "exactly calculated" number as defined by him/her earlier. Thus, his/her statement,"The value of pi can never be known exactly, so surface areas of spheres cannot be calculated exactly." is blatantly false. The algebraic relationship between numbers still holds.
    Last edited: Dec 2, 2011
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook