First of all, they used Transit Time of a boat on a water stream as the analogy. One can see quite easily, that waves are wavelength compressible/extendible, by simply visiting a local river where it empties into a lake. It is easiest to see if the waves are coming in off the lake and climbing the counterflowing river.

Boats are singular entities and not at all compressible in length (regardless of
the mathematical thrashing Lorentz gave them). Waves come in trains, and certainly are compressible/extendible.

And finally, it is not the Transit Time that is the critical criterion, but rather; the total number of waveforms on the 180-degree folded paths. If the total number remains constant regardless of setup rotation, then the fringes indication remains static.

It turns out that in a headwind (aether wind versus lightwaves), the number of waveforms added to one leg is exactly the same number as the waveforms subtracted from the other.

I used Relativity of Velocities and Relativity of Position-change to prove it, along with some simple arithmetic. See my Drawing 64.

In the evaluation of the MM Exp result, they never took wavelength c/e into consideration, as is with the fellow in the link to "Real Science...".

If anyone wants to demo wave dispersion to himself, it is a simple matter of tossing a stone into a local reasonably smooth stream or pond.