I remember the way that it was taught when I was taking Multivariable Calculus was to parameterize the curve and vector field. This also seems to be the method I've found online.
However, in my current class to solve such a problem, we simply take the dot product of the vector field with the differential length, which if it's in rectangular coordinates is: xdx+ydy+z*dz.
What's the difference between the two methods? Why do you need to unnecessarily parameterize everything if you can simply use the second method?