Georgina,
It takes 180 degrees not 90 to reflect back along the arrival axis. But if you read my specification more carefully you'll see the rotation of the 2nd mirror is also around the BEAM axis. That is, around the "axis" of say a laser beam arriving at a 57.2 degree incident angle. The beam will then STILL arrive at the same incident angle but now be reflected 'SIDEWAYS' to your eye.
Except that it wont. The beam entirely disappears. There is no interaction with the incident light! ('time reversal' among other nonsense can explain it as the light didn't then arrive!)
Past analysis, i.e. Ference 1956, suggests this proves light has ONLY lateral 'waves', inconsistent with almost all other physics! (The first mirrors glass absorbs (refracts) y axis light, (so it's 'plane polarised), the second absorbs the z axis light, so there's none left, so there can be no longitudinal (x axis) wave component!.
However, by using the same assumptions as the DFM a consistent logic is possible, where oscillations can retain the full 3D spherical freedoms found elsewhere. Longitudinal components are simply absorbed or reflected by proportion to angle.
The 'moving mirror paradox' is akin to the Dynamic Casimir Effect, itself apparently fatal to the 'Wonderland' interpretation of SR. It's logical resolution uses the same foundations. Maxwell's near/far field transition is simply a speed and wavelength change at the 'two-fluid' plasmon surface fine structure. 'Virtual' electrons are REAL! just in the 'other' frame, continually propagating and annihilating (a well known process at the Debye length).
Light hitting the mirror moving in the vacuum is then modulated to c in the mirror frame on arrival, reflected, then instantly modulated back to c in the ('far field') 'vacuum' frame. At an antenna and for longer wavelengths this 'transition zone' is well known to vary subject to lambda.
That means you're right about subjectivity. At a lens light's speed and wavelength are modulated to local c by each observer's lens TZ, giving the 'personal' Doppler shift found. Quite simple really, just a bit different to some current assumptions. See the logic?
Best wishes
Peter