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

Confusion about notation and convention regarding differentials as objects and operators

  1. Jun 9, 2015 #1
    I am currently reading "Differential Equatons with Applications" by Ritger and Rose, and I need some clarification about some notation and convention that they are using. I think it all stems from a lack of clarity of the difference between the operator d/dx and the "object" (I dont know what to call it, I dont really know what it is) dx

    If I have an equation of the following form

    d(f(x,y)) = h(x,y)dx

    the approach they then take is to "integrate both sides" and they remain with the following equation

    f(x,y) = g(x,y) + c where g(x,y) = ∫h(x,y)dx

    the integration on the right is clearly with respect to x, I am not so clear as to what they are doing on the left however. It seems like they are asserting that the integral of the derivative of a function is the function itself. But with respect to what are they integrating it? Is this only with respect to x? If so then the answer is wrong since ∫ d(f(x,y))dx is in general not f(x,y) . Is this some sort of line integral and we are using gradient theorem? If so are we doing the same thing on the other side of the equation? I dont think so, as we are only integrating with respect to x

    The other confusion comes from the following, if have the same initial equation as above, d(f(x,y)) = h(x,y)dx is that the same as (d/dx)(f(x,y)) = h(x,y)? If yes, how so? How about if I have something of the following form ∂/∂x (f(x,y)) = h(x,y)?

  2. jcsd
  3. Jun 9, 2015 #2


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    I'm not sure that is meaningful in isolation. It certainly looks like it might mean ∂/∂x (f(x,y)) = h(x,y), but I can imagine other circumstances. E.g. we might be considering stepping a small distance along a path in the XY plane, so that in general d(f(x,y)) = h(x,y)dx+k(x,y).dy, but that it happens in this case that k(x,y) (at some point) is zero.
    Assuming it does mean ∂/∂x (f(x,y)) = h(x,y), we can integrate wrt x to get f(x,y)=∫h(x,y).dx+k(y), where k is a 'function of integration', analogous to a constant of integration.
    ∫ d(f(x,y))dx does not parse. You mean just ∫ d(f(x,y)).
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook