Dale said:
Sure, it fixes the coordinate system, but fixing the coordinate system doesn’t make it independent of the coordinate system.
I never said ”energy” was coordinate system independent. I said ”energy in coordinate system S” is coordinate independent. The latter fixes the system in which energy is to be found. You can find that energy from computations made ising any other coordinate system and it will still be the same energy.
Dale said:
Similarly, the tension in a spring depends on the length. You can fix the length, but that doesn’t make the tension independent of the length.
Again, I never claimed so.
Dale said:
When you fix the coordinate system, a quantity that is coordinate dependent becomes well-defined, not coordinate-independent. I think you are confounding well-defined with coordinate-independent.
I disagree. If you have a set of basis vectors, the coefficients of any other vector in terms of that basis are scalar quantities. They are literally found by the inmer product of the (dual) basis and the vector. It does not matter whether I compute that inner product using coordinates that are based on the basis or a different coordinate system. However, the coefficients are quite obviously going to depend on the basis you use for expressing the vector.
Dale said:
I am not completely sure which quantity you are referring to by “energy in coordinate system S”. I think it is et⋅p, which depends on t.
Indeed, and t is still a scalar function in any other coordinate system. It just does not have the expression ##t = t’##.
Dale said:
It is well defined and, as you say, can be calculated in any other frame. But none of that removes the fact that the coordinate t is literally right there in the formula and changes when you change coordinates.
Yes, t is there, but it is not equal to t’. It is ##t = \gamma(t’ + vx’)##. This is just the regular inverse Lorentz transformation.
Obviously ##e_t \cdot p## need not be equal to ##e_t’ \cdot p##.
Let us take it back to Euclidean geometry in 2D. Given an orthonormal basis ##\{\vec e_i\}## you can express any vector ##\vec v## as ##\vec v = v^1\vec e_1 + v^2 \vec e_2##. The quantities ##v^i## do not depend on whatever coordinates you pick. You can choose to use the Cartesian coordinates based on ##\{\vec e_i\}## or you can use Cartesian coordinates based* on ##\{\vec e_i’\}## (or any other curvilinear or oblique coordinates), the ##v^i## will still be given by ##\vec e_i \cdot \vec v##.
* By this we mean that the coordinates of a point ##p## are the ##x^i## in ##\vec x = x^i \vec e_i## with ##\vec x## being the position vector relative to some chosen origin.