Re: twin clock problem - SR experts help!


Bilge wrote:
> Harald EPFL:
> >"Dirk Van de moortel" <dirkvandemoortel@xxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote
> >in message news:wfEPe.11208$qc2.3178@xxxxxxxxxxxxxxxxxxx
>
> >> When you are alone in the desert without an atomic clock or
> >> wrist watch or sun or stars, you can define time and "the passage
> >> of time" with your heart beat, and do physics and engineering.
> >> You don't need reference standards. You need a clock,
> >> something that does pom pom pom pom pom pom.
> >> You count the poms and call it time.
> >> When you have more than one kind of clock, you can
> >> compare one against the other and compare the times you
> >> define and read on those clocks.
> >
> >Sure. And when you discover that they don't indicate the same, you will have
> >some choosing to do, exactly as I desribed above...
>
> I'm afraid that I don't see why anyone would choose to use an
> ``incorrectly'' calibrated clock for a thought experiment, or why
> you can't just specify a ``thought calibration'' which illustrates
> whatever preconception about time keeping you want to impose. I think
> the basic idea behind the idea that ``time is what a clock measures,''
> is that it doesn't specify a particular kind of clock that would
> depend upon a detailed understanding of how the clock is supposed
> to work. Having a lot of things called clocks is not a useful way
> to go about it, unless you have some reason to believe the clocks
> actually measure the same thing. That requires a physical model
> for each clock.
>
> >> If you can define regularity or "correct calibration" without
> >> using the concept of time, then I agree with your suggestion
> >> that "a correctly *calibrated* clock defines the passage of
> >> time"
> >
> >Hey Dirk, "temperature" is *not* what any alcohol thermometer indicates;
>
> Sure it is, except perhaps in the world of management in which managers
> and marketeers like to believe whichever thermometer they can buy to reads
> the best temperature for the advertising they envisioned. An alchohol
> thermometer is an alcohol thermometer. If it has no numbers on it, its not
> difficult to put it in some boiling water and scribble a number on it,
> then put it in some water where the triple point is observed and then
> divide up the interval into equally spaced divisions, which can be checked
> to see if the divisions coincide with what you think temperature is
> supposed to mean. How ``good'' it is, amounts to how well the thermometer
> reads what your scribble marks say it should read under the same
> conditions you used to decide where to scribble.
>
> A ``calibrated thermometer'' is the same thermometer, except that
> someone else does the scribbling and a disclai^H^H^H^H^H^H^Hcertificate
> comes with it that tells you the extent to which the manufacturer will
> will sympathize with you if you trust it for something important. In
> any real experiment, if a calibration matters, the experimentor does
> it himself, since he's the one whose reputation for doing careful
> work is on the line.
>
> >evenso "time" is not affected by your emotions while your heart beat is
> >affected by that!
>
> That doesn't matter unless you are completely incompetent as an
> experimentalist. Experimentalists have to live with imperfect
> and even crummy instruments. Clever experimentalists try to construct
> experiments that don't rely to much on knowing what an instrument
> reads relative to some NIST calibration standard. Instead they try
> to design experiments that mainly depend on knowing that the
> instruments read the same thing consistently, regardless of what
> absolute number that might be. Ideally, one tries not to count on
> that too much if possible.
>
> For example, if you want to divide a bag of sand into some
> number of equal piles, then you'll do better using a balance
> that doesn't tell much of anything at all, but which can
> be nulled out, than you will do using a calbrated scale to weigh
> out each pile. (I mean, your average cocaine dealer is not the
> best experimentalist in the world, but the measuring instrument
> of choice is ye-olde triple beam balance, because it works
> reliably under all sorts of conditions, many of which I imagine
> are much less ideal than most regular labs.)
>
> >I already imagine your time go quicker when you see a
> >pretty girl - pom pom pom ;-)
>
> >We can use such instruments to do experiments with limited precision.
>
> Precision is a statistical quantity. It depends on how well
> your instrument reproduces a number that you _think_ should
> be the same and how much that uncertainty matters for the
> length of the interval you are measuring. If you calibrate
> your heartrate using a thousand trials for the same measurement
> of something that lasts about 100 heartbeats, then you ought to
> get excellent precision for a measurement that takes 1 million
> heartbeats. If you madethree such measurements that were
> different by a thouand heartbeats, you'd still do better than
> a tenth of a percent.
>
> >The regularity of any measurement standard is verified by for example
> >comparing it with other devices and processes, just as it was discovered
> >that an alcohol thermometer is less good than a mercury thermometer.
>
> Discovering which is better depends upon how well each thermometer
> yields the same result when measuring the same thing. Seeing if
> two thermometers give the same result for some arbitrary measurement,
> can only tell you that one of them is wrong, not which one is wrong.
> Which is wrong requires specifying a definition for ``right.''
>
> >By comparing an abundance of instruments one can determine which ones are
> >less regular under what circumstances and in what intervals.
>
> No, you can't. All you can determine is that the instruments don't all
> give the same result. You can't even tell if any of them are ``regular''
> much less which ones are more ``regular'' than others. Before you
> can say anything about that, you have to define ``regular.'' The only
> way to do that is to choose some natural phenomenon as a definition,
> based on some model that you have some reason to believe applies
> the way you think it does.
>
> The reason that atomic clocks are believed to be so precise is
> because their precision depends entirely upon the statistics of
> a random process, which happens to be precisely how quantum
> mechanics describes atomic transitions. If any particular atomic
> transition occurs probabilistically (i.e., has no cause), then
> such a clock is as theoretically perfect as any clock could
> ever be, because it doesn't depend upon any mechanism. It only
> depends upon statistics.

Atomic clocks don't rely on statistics... idiots who don't
understand how they work do.

<<
where

m is the mass of an electron
mp is the mass of a proton
a is the fine structure constant (1/137.036)
c is the speed of light.
This is a much smaller perturbation than the
fine structure or Lamb shift. >>
http://en.wikipedia.org/wiki/Hyperfine_transition

Sue...

.

Relevant Pages

  • Re: twin clock problem - SR experts help!
    ... You need a clock, ... thermometer is an alcohol thermometer. ... instruments read the same thing consistently, ... your heartrate using a thousand trials for the same measurement ...
    (sci.physics.relativity)
  • Re: twin clock problem - SR experts help!
    ... You need a clock, ... > thermometer is an alcohol thermometer. ... > instruments read the same thing consistently, ... standards are as small as 400 cm3), ...
    (sci.physics.relativity)
  • Re: Math is Easy, Physics is Hard
    ... Temperature is a measurement ... of energy flow between two systems put in thermal contact. ... of a thermometer, ... If you see a clock, ...
    (sci.physics.relativity)
  • Re: time to try brisket again
    ... Will be doing the thermometer & fork tests ... as opposed to the clock. ... And my smoker runs at about 200-225 degs ...
    (alt.food.barbecue)
  • Re: Circular motion in SR
    ... transitions of cesium isotope molecules, ... the only measurement of time allowed. ... c in S' as measured by a clock in S'. ...
    (sci.physics.relativity)
Quantcast