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

Differentials and Errors

  1. Oct 26, 2011 #1
    I asked a question a few weeks ago about 'splitting' the derivative. The thread can be found https://www.physicsforums.com/showthread.php?p=3581188#post3581188"

    The answer to why it can not be split is because dx does not exist, it is simple a notation and not a fraction.

    However, I just started Differentials and Errors, and the paper I read said.

    The ratio of the two derivatives is actually the derivative of the function.

    [itex]f'(x)=\frac{dy}{dx}[/itex]

    and the relationship between the two differentials can be given by

    [itex]dy=f'(x)dx[/itex]

    Is this not 'splitting' the derivative?

    Thank you
     
    Last edited by a moderator: Apr 26, 2017
  2. jcsd
  3. Oct 26, 2011 #2

    lurflurf

    User Avatar
    Homework Helper

    dx does exist and dy=f′(x)dx is fine

    The actual definitions in use can very and sometimes confusion can arise. Often the implicit function theorem and its conditions are in play.
     
  4. Oct 27, 2011 #3
    I think you mean the ratio of two differentials , yes you are right , dy and dx are merely notations , and don't exist seperately but when you write [itex]dy=f'(x)dx[/itex] , dy becomes function of variables x and dx , and meaningful when used together as in the case of dy/dx , also see the following link
    http://en.wikipedia.org/wiki/Differential_of_a_function" [Broken]
     
    Last edited by a moderator: May 5, 2017
  5. Oct 27, 2011 #4

    lurflurf

    User Avatar
    Homework Helper

    That is like saying 3/4 is merely a notation and 3 and 4 do not exist separately.
    The advantage of differentials is when we have several variable we can understand how each changes with the others independent of the representation as differentials are invariant. dy/dx means the same thing if we have
    y=f(x)
    x=g(y)
    h(x,y)=0
    w(u(x,y),v(x,y))=0
    and so on
    it would be quite limiting to say dy/dx=f'(x) the end, if we change variables we will start from scratch.
     
  6. Oct 27, 2011 #5
    well said ! , But i don't say dy/dx is merely a notation , tell me what's dx ? ( without talking anything about dy)
     
  7. Oct 27, 2011 #6

    lurflurf

    User Avatar
    Homework Helper

    Again different definitions of differential are used in different contexts so this is for the purpose of regular calculus
    dx is just some variable for an arbitrary change in x
    which is not the most interesting part
    given some sufficiently well-behaved curve or relationship of variables we can define a tangent space that depends upon the point and the curve at that point, but in no way depends upon specific coordinate choices. The projection of the curve is a linear function of the differentials and approximates the curve well near the point.
    thus
    f(x+dx)-f(x) is some function of x and dx
    and
    f'(x)dx is some function of x and dx
    they do not agree in general only approximately near x

    This only works so well for total differentials of first order.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Differentials and Errors
  1. Differentiate this (Replies: 2)

  2. The differential of (Replies: 2)

  3. Error estimate (Replies: 0)

  4. Differentiable (Replies: 2)

Loading...