Edwin,
Nice essay. Fun format and moot points to your usual excellent standard (though AE didn't get much of a word in!) Of course I agree local 'ether' frames, but Einstein didn't get a chance (or forgot) to mention his key final concepts (1952 'Addendum'). Which I agree, and relate to your p5 'boxes'. Please advise what's wrong with this;
1. First do away with the 1st (bottom left) box. If at rest in the ambient frame it's 'part of it'.
2. As the waves reach the top box they're absorbed by the boundary fermions, re-quantized and re-emitted at local c in the new rest frame (now in Maxwell, Lorentz and radio engineers Transition Zone {TZ} and 'far' field). They are thus Doppler shifted to red. If they're sequential time signals? - they thus appear to be slowed (dilated) ..or 'contracted' (blue shifted, 'sped up', if moving the other way).
3. If the lower box were also in some other state of motion that TZ process would also happen as it entered the ambient medium. (There would then be 3 local speeds c visible by displacement by the reader!!! though all signals do c locally.)
4. In the case of the train, light within the train does c in the local train rest frame and on exit Doppler shifts to do c in the track frame. (We can substitute Earth and its ionosphere or any lens, for the box & train). Watch a pulse through the passing windows and it looks like c+v.
5. The TZ is from (we know) a micron deep (at prisms & lenses) to a few parsecs for galaxy clusters, the more diffuse the medium the more gentle the curvature (& birefringence) and it's also lambda dependent (antenna engineers know all this). There's a short low quality video of a moving box if you'd like it.
That's based on the discrete field model, (DFM) consistent with what Einstein forgot he'd said in '52 (too much to drink I expect!) describing inertial systems as "spaces with spaces, not thought of as bounded, in relative motion". I (now) don't find that "non-intuitive" at all!. It seems consistent with all evidence I can find & seems to resolves the issues. That's subtly but importantly different to Hertz's view. Can you identify where and why it may appear flawed. or raise any questions?
The classical QM solution (as Declan employs) emerges from the same interaction model (though not just 2 states & 3 axes, and also including Cos2,- see Declans string).
Great essay whatever. Another top score due I think!
Very best
Peter