[ntp:questions] Re: Extracted fom comp.risks - What Time Is It?

David Schwartz davids at webmaster.com
Mon Aug 18 19:26:49 UTC 2003


"H. Peter Anvin" <hpa at zytor.com> wrote in message
news:bhpo89$k70$1 at cesium.transmeta.com...

> I think the big question is how do you *synchronize* these various
> clocks.

    That's definitely the harder part.

> An atomic clock tells you what time it is in one particular location.
> Any kind you move a clock you have to deal with time dilation issues,
> and when you communicate its time using signals you have propagation
> issues.

    You can either ignore the time dilation issues or correct for them if
needed. It used to be pretty common to fly a synchronized atomic clock from
one place to another to synchronize the atomic clock at the destination. GPS
has pretty much replaced this (painful) process.

> To make things even more complex, the speed of time actually
> varies with the strength of gravity and therefore varies with
> altitude, latitude and geological composition, although I *believe*
> this is currently below the threshold of measurability (time dilation
> due to moving an atomic clock is, however, very much measurable and
> can amount to tens or even hundreds of nanoseconds on a
> round-the-world journey.)

    I don't believe anyone corrects for gravitational effects. Time dilation
definitely requires compensation when you have to fly an atomic clock from a
standards lab to your radiotelescope observatory.

    DS





More information about the questions mailing list