FrediFizzx wrote:
I like the way Griffiths described an inertial frame in "Introduction to
Electrodynamics".

"To avoid this trap we define an inertial frame formally as one in which
Newton's first law holds. If you want to know whether you're in an
inertial
frame, throw some rocks around---if they travel in straight lines at
constant speed, you've got yourself an inertial frame, and any frame
moving
at constant velocity with respect to you will be another inertial frame."

I don't see that it has to be any more complicated than this.


I like Landaus definition - it is a frame is which space and time are
homogeneous and space is isotropic. It is equivalent to the one above via
the POR (nice little exercise to show this). Problems come when you look at
it really closely. First you must specify a coordinate system. Physically
how do you do this - via rigid rods (SR later shows they do not exist) or
exactly how? Secondly how do we synchronize clocks so we have a universal
time? Using OWLS? Then the speed of light is constant by definition. Take
two clocks at the same point then sync them and slowly move them apart? Are
you sure they will then stay synced? Via two way light speed? Only if your
really sure of the isotropy property. Which then begs the question how do
you show a frame of reference is homogeneous and isotropic? Although I am
no expert he it just seems like a total mine field to me that I just hope
someone has sorted out.

My solution is you accept these things as extra axioms ie you can construct
from stationary rods that are rigid in the classical domain Cartesian
coordinates (guaranteed by Euclidian geometry which you assume valid for
stationary or slow moving lines and points). You also assume you
conceptually have a synced set of stationary clocks at all points. But
physically it still seems a problem.

Thanks
Bill