Daryl,
Thanks. I don't think the low scores are anything personal. I'm sure I've had perhaps a dozen 1 scores, as many others. It may be indicative of the integrity of many in science. As effectively a 'professional' over 3 fields I can see the differences, but I've also seen the honesty and integrity in most, certainly yourself. I've just rated yours, taking it closer to it's deserved level.
You say my derivation; "doesn't seem right". That's good! As Feynman predicted; the 'correct' answer cannot first look right as it must be different!
Now let's study that rubber band, because it is THAT which we are here analysing! In current 'interpretation' of SR it's just a hand waving effect with some manipulated symbols to maths it, but no real mechanism. (Remember it's only in an interpretation NOT the theory itself, which as Einstein made very clear in 1952, is; "entirely contained within the postulates."
I now invoke the bit you've agreed, that light representing an 'event', once emitted, is just physical evidence 'at large' so liable to Doppler and other effects on refraction (which is change of propagation speed in a dielectric medium). What changes then is wavelength lambda (the first wave is slowed before the second) and, similarly the 'distance' between the flashes.
How is that changed? Because each 'flash' is absorbed by the matter in the new FRAME and then re-emitted at c in that frame. That is an effect additional to any refractive index difference between the media. It is also precisely what is found in optical sciences (as all experiments inc. Fizeau, Sagnac, Wang etc. etc.) That is also what gives the 'nonliear optics' effects not presently explainable theoretically.
Now of course it's unfamiliar to you. But it's entirely consistent with all observation, and indeed resolves paradoxes (i.e.; recovers Snell's Law and Fresnel refraction at Maxwell's near/far field transition zone, also explains KRR and the Kerr effects). So lets test it again;
Flash C1 is emitted IN the front of the train. The train and air represent a single 'inertial system' through which light propagates at c, (so a 'discrete field' = DFM). the observer also at rest, so the event remains 1ms, and with no Doppler wavelength change.
Flash C2 from the light on the side of engine is outside the train but in the same inertial FRAME (at rest with the observer). Again no delta lambda on observer interaction ("detection"). However, it arrives BEFORE C1! This is because is DID change speed to propagate at c in the frame of the air outside the train, but then changed back on re-entering the original (emitter/detector) frame. That surprisingly is as found and as SR.
Flash C3 (from a fixed post by the track) ALSO arrives before C1, so with C2, as it also propagates at c in the outside air frame. It then also shifts on meeting the observer frame, but this time there had been no INITIAL shift, so it is found to be blue shifted; i.e. both wavelength lambda AND the 'space-time period' have undergone 'length contraction'. (A flash from behind would be dilated or red shifted).
Shocking at first I know, but not for long, and none the less far more consistent that current incomplete interpretation, and deriving SR direct from the quantum mechanism invoked (= unification). Just imagine the elephant in the room ~200 times as big as you were expecting and you should then be able to make it out. The elements in my essay are just a few of the rather powerful consequences. You may recall others from last years.
Peter