You criticized those who demonstrate circuits that can't be modeled. I asked first. What model are you talking about? Perhaps those people are trying to show you that it's your model that is broken. Not their circuits of their demonstrations.
You lost me somewhere. I didn't say the circuits couldn't be modelled, they can always be modelled but whether the models are correct or not is another matter. My criticism was of using physically impossible or experimentally impractical circuits to 'demonstrate' a theory. You can model those all you like, but you can't test your model in the real universe. I proposed no specific model for this case, others came up with the transmission line. I think there is more to it, as even with no resistance it should also radiate energy into space--so even an ideal transmission line is not a perfect model. But that doesn't affect the outcome of the question posed.
Correct my donkey. In LT-Spice, currents flow immediately between the terminals of a capacitor giving a spurious spike before 1 m /c seconds. There is no simultaneity in the universe. This is what Einstein deduced from Maxwell's equations when he formulated his theory of special relativity.
Numerous people, myself included, have readily observed those shortcomings in the lumped-element approximation of the transmission line and even your cretinous bete noire specifically pointed it out as such. We all assume that any phenomena is limited to light speed. Nobody (at least not me) ever thought there would be a faster-than-light spike. By correct, I mean the stepped transitions of indeterminate height, indeterminate because we don't know the impedance of the light bulb. And I think those better models, which seem to match the experiments, were done using an explicit transmission line not lumped elements.
Relativity is one reason that Derek's proposed correct answer is clearly wrong. If you accept his impossible magical conditions, an observer at the switch will see the light turn on at about 6.66ns, or twice as long as he says. An observer at the light will observe both events to be nearly simultaneous. Only an observer halfway between the two or very far away above or below the plane of the wires will see the approximately 3.33ns transition time. So (E) None of the above--unless you are in just the right spot.
The correct modeling of this problem can only be precisely predicted using the Poynting theorem, which can only be deduced from Maxwell's equations. This is why engineers study these "unecessary" things in their respective degrees.
OK, then I'm willing to learn. If you can, specify a reasonably buildable scaled down version of Derek's impossible circuit, using real components. If you then model it using Poynting's theorem for us, we can then model it in LT Spice and compare the results. If those results differ significantly, I'll build it experimentally if I can. I have wire, sawhorses, some old 300-ohm twin lead (I'm not sure how much), scopes that can resolve about 1ns and I'll have to work on a fast edge, long period square wave. It needs to fit in my garage, or worst case in my backyard, so not too big.
What do you have against audience engagement?
Nothing. Look at this whole discussion!