> > Suppose I'm at A. I send a signal to B, one light-year away. Then I go
> > on walkabout, speed
> > unspecified, and return to A, stop and eat lunch, now at rest relative
> > to B. The signal returns.
> > When did my watch speed up so that it records exactly 2 years for the
> > signal to return?
>
> Your watch does not record exactly 2 years.

Oh... I thought the "time" for light to travel from A to B was equal to
the "time"
it took to travel back again from B to A, that's one year each way, my
watch was at rest at A
at the start of the experiment, is at A once again at the end, again at
rest, so why doesn't it record two years?
Seems to me there must be something wrong with it if it doesn't. Good
grief, it would indicate the light had travelled from A to B and back
again faster than c if it showed less than 2 years, wouldn't it?
Can't have that, can we? It always travels at c in MY frame of
reference, doesn't it?

> Your watch will record less than two years.

Really? Hmmm... assertion carries no weight, y'know.
You'll have to use some hard sums to show me how.

> The watch you left behind at A records two years,

Ah, yes, the clock I left behind was a light clock, you see.
The alarm was set for 2 years. It works by sending light a distance of
1 light year to a mirror. When the light is reflected and returns it
goes through a photomultiplier, produces an amplified signal and that
is then used to trigger the ringing of a bell. I designed it myself.
And I'm sure you are right, it does record two years. Cute, huh? And
I'm sure I calibrated my wristwatch to it correctly.

> provided A and B remained in some inertial frame
> at rest w.r.t. each other.

Yeah, sure, but I was tied to that same inertial frame at the start and
at the end of the experiment, wasn't I?
Does going down to the pub for a pint and coming back before the bell
rings invalidate it, then?
AD