> This may be an old thread, and this post may not add to it much, but it's relativity.
>
> The Train Experiment
> --------------------
>
> There are two men. Each has a light clock. A light clock simply consists of two mirrors facing each other, with a single pulse of light moving between them:
>
> //////////
> -----
>
>
>
> -----
> //////////
>
> The top mirror has a light sensor and emitter. It sends a pulse of light down to the bottom mirror, then the sensor catches it when it returns. The distance is fixed, as is the speed of the light, so therefore the time between sending and receiving is also fixed.
>
> Each of the men has one of these clocks. They look at each other's readout and see that the clocks are in sync. The period would be the same for each clock.
>
> Now one man gets on a train. Himself and his clock are still clearly visible to the other man, who is standing at the side of the tracks. They look at each other's readouts, and what do they see?
>
> Taking the frame of reference to be man not on the train. His clock will still tick at the same regularity as it always did. Nothing has changed. Then he looks at the other man's clock.
>
> It is moving from west to east. Following the path of his light, he sees:
>
> --------------------------------
> \ / \ / \ / \ / \ / \ / \ / \ / and so on.
> --------------------------------
>
> The light has to travel a longer distance (part of it laterally) in its oscillation. It does this at the speed of light. Therefore, it takes longer. The clock on the train will count slower than his own.
>
> Exactly the same principle holds if you take the man on the train to be stationary. So, both of them see the other's clock as being slow. Who is right?
>
> The Whole Story
>
> The Beginning
>
> They're both standing at the station, their clocks in sync. One gets on the train. It starts to accelerate. Now, this is where the problem lies. It accelerates. It doesn't just jump to a new velocity. Anyway, during this acceleration, the clocks start to experience the relativistic effects. Both of them see each other's clock start to slow down.
>
> The Middle
>
> Things are running smoothly. Each clock has a constant period to each observer.
>
> The End
>
> He decelerates. Each clock appears to "speed back up" until they are running the same period again. Which clock counted the most ticks? I don't know.
> There are a few fundamental flaws in this situation:
>
> 1) The distance between the men. This cannot be constant, otherwise the train would have to be going around a curve, which is not an inertial frame of reference. So, let's say the man was watching from afar, like so:
>
> (overhead view) . . . X
> /- The traveller
> v
>
> [X ]========================================[ ]
>
> In this situation, the traveller is accelerating with respect to the observer. He starts off almost coming straight at him, and ends up almost going directly away.
>
> So what's the result? I don't know. There has been acceleration, and so no inertial frame of reference, in this scenario. Ah well. I've got too much of a headache to continue right now, and I don't even know if anyone will read this far. Maybe Dave, gremlinn, or Wes could shed light on the situation? I haven't argued physics for a long time.
>
> Tub"I can't believe I drew the ASCII stations"ba

I think that when they are separated and the train has stopped (hence rejoining the other guy's reference frame), you'll find that the train guy's clock has made fewer ticks. You can't argue the problem the same way using the train as a reference frame, even though the standing guy appears to accelerate, because the train isn't an inertial frame of reference.

How much behind will the train guy's clock be? Exactly half the amount that it would be if he made the return trip (because you could duplicate the periods of acceleration/deceleration/coasting and get the same time difference.)

Why will the train guy be dropping further and further behind? I don't know how to work the math out at the moment (it's been too long since I learned this) but the whole round trip is just the Twin "Paradox", which you can find info on at the link I am providing (maybe there are better explanations; this is just the first I found).