"length contraction" according to Lorentz and Poincare - An example

9 views
Skip to first unread message

Harry

unread,
Dec 19, 2004, 2:24:17 PM12/19/04
to
In a recent posting I saw the claim:

"Under SR, A thinks that B's ruler is shorter. B thinks that A's
ruler is shorter.
Under LET, A thinks that B's ruler is shorter. B thinks that A's
ruler is *longer*."

"LET" referred to the mathematical physics as found in Lorentz' 1904
paper, and "SR" referred to the mathematical physics of as found in
Einstein's 1905 paper.

I don't think that that is correct; in any case, Lorentz obviously
didn't feel that he should be stuck to some little errors and
incomplete understanding that he had when writing a certain paper, and
neither did Einstein. After 1905 Lorentz taught relativity and he did
not propose a separate "LET" as there is no difference as far as
mathematical or experimental physics is concerned.

That reciprocity of observations directly follows from Lorentz'length
contraction and clock slowdown is not immediately obvious.

Recently I wrote for another newsgroup an example of the
Lorentz-Poincare interpretation of "length contraction" and "time
dilation" without making use of the Lorentz transformations.

I copied it below as it may help some with intuitive understanding
("feeling").
_____________________________________________________________________________


EXAMPLE OF APPARENT MUTUAL LENGTH CONTRACTION:
A MOVING RULER THAT PASSES A STATIONARY ROD

Text books show that the inverse Lorentz transformation follows from
the Lorentz transformation. However, abstract mathematical evidence
does not fully satisfy everyone, and the philosophical interpretations
are open for debate. It may appear impossible that with a shortened
ruler one could measure other objects to be shortened.

It is helpful to understand the possible mechanism of the phenomena. I
will try to transmit here the understanding thanks to the explanation
according to Lorentz, by only using his "length contraction"
(contraction of objects) and "local time" (clock slowdown), in
combination with Poincare's calibration procedure.

It should be kept in mind that Lorentz combined the "absolute space"
concept of Newton's physics (used for observer-independent inertia)
with the "ether" concept of Maxwell (used as a carrier for waves);
what follows should not be confused with Einstein's interpretation of
relativity or the Einstein-Minkowski's Space-Time concept of
relativity.


THE EXAMPLE:

If a steadily moving system has two clocks, each at one end of a
ruler, and the velocity relative to the stationary ether is v, then,
according to Lorentz-Poincare's theory,

its length L = L0 / gamma,

with gamma = 1/(SQRT(1-v^2/c^2), L the length of the ruler in
movement, L0 its length in rest. Clock rates are expected to reduce by
the same factor.
We will assume that this can be measured with a system of coordinates
that is itself resting in the ether. (In practice only effects on
clocks have been directly measured, although MMX was explained with
length contraction.) Only in the ether frame is light speed truly
isotropic, i.e. c in all directions.

For simplicity, for this example we will use a system of reference for
defining durations and lengths that is resting in the ether or
"absolute space", so that v is the "absolute speed" (= relative to
"absolute space"). we'll also only consider one dimension. Thus by
definition only in this frame will duration and length measurements
provide the true values.

Let's take v=0.8c, so that gamma = 5/3. Thus according to the theory,
the ruler will be contracted to 3/5 of its rest length: for this
example we'll assume a ruler that is 6 m in rest, so it will be 3.6 m
long. Similarly, the clocks will tick at 3/5 of the rate in rest,
making one tick 0.6 s of duration. Because this is determined relative
to the stationary ether or absolute space, the effects correspond to
physical reality.

Now, what length will be measured with the moving system, when it
passes a 6 m long rod that is in rest?

According to the theory, the measured relative speed is the same: the
velocity v'= -v in standard notation, or, what I prefer for
mathematical symmetry: defining approaching speeds as positive, v'=v.
(Primed symbols are the values as determined with the moving reference
system.)

That the perceived speeds are equal can be demonstrated as follows.


V'=V:

If the clocks have been synchronized with light or radio pulses, the
synchronization proceeded for example as follows, with a signal sent
from the middle towards the clocks:

Light signal to both clocks: assumed by convention to take 3[m]/c[m/s]
or 3/3E8 = 10 ns.
Note the convention: those observers, ignorant of the speed of their
system, use as working hypothesis that light speed is isotropically c
relative to their chosen reference system.

However, in reality and as noticed from the rest system,
Light signal to front clock:
1.8 [m] / {(1-0.8)*c} [m/s] = 9/c s
Light signal to rear clock:
1.8 [m] / {(1+0.8)*c} [m/s] = 1/c s
------ +
Synchronization error : 8/c s = 26.7 ns

Thus the number of ticks that C2 is ahead of C1 is:
t2'_offset = 3/5 * 8/c = 4.8/c ticks = +16E-9 ticks.

If we define t'=0 at the moment that clock C1 meets the rear end of
the rod, that will be indisputably so for C1 : t1'=0.

C2 C1
ruler ---------------------- --> v
rod______________________________________

However, from the system in rest it will be noted that at that moment
C2 already indicates 16 ns.
C2 will meet the front of the rod when the 3.6 m ruler has reached it,
moving at 0.8c, thus after 3.6/(0.8*c) = 4.5/c s. C2, ticking slow,
will then have advanced by:

delta_t'= 3/5 * 4.5/c = 2.7/c ticks or 9E-9 ticks.
Thus in total:
t2'= t2'_offset + delta_t' = 4.8/c + 2.7/c = 7.5/c ticks or 25E-9
tick.

Therefore, according to speed determination in the moving system, its
apparently 6 m long ruler was passed in 25 apparent ns, or at 0.8c, so
that the speed measurement with the erroneous instruments still
provides the correct result.


THE LENGTH MEASUREMENT:

Now we are ready to determine the length of the stationary rod
according to the moving system:

1. If the length is determined with a clock, based on the measured
speed, obviously the slower tick rate causes the same underestimation
of the rod's length: it counts only 3/5*25 = 15 nanoticks for passing
the rod, so that the rod's length seems to be:
0.8c [m'/s'] * 15E-9 [s'] = 3.6 m'.

2. More complicated is it if we compare lengths. For that we need to
determine at what point of the ruler t'=0 when the rod's left end
passes it.

C2 C1
-------------------- --> v
rod______________________________________


We established that t2'=0 at a time 8/c= 26.7 ns earlier, thus when
the clock is 8/c * 0.8c = 6.4 m to the left. Funny enough, C1 has then
not yet met up with the rod, but the time as determined from C1 is
then -26.7 ns so that isn't relevant. What matters is that the front
end of the rod has not met with C2 when the clock time is 0, so that
the rod seems shorter than the ruler. In principle the mechanism is
thus explained.

For a precise determination one needs to either send a light signal
from C2 to the rod and back at t2=0, or (for the calculation less
complicated!) one needs to have clocks everywhere and then the clock
that reaches t'=0 when the front end of the rod passes it will
establish the apparent length.

In theory that should be at point: x'= 3/5*6 [m] = 3.6 marks (a mark
is an apparent meter). The clock at that point would have an offset as
follows: 3/5* 16E-9 ticks = +9.6 nanoticks.
That point is 3/5*x'= 2.16 m to the left of C1. Thus the ruler has
then 6-2.16 m= 3.84 m to go before C1 meets up, or 3.84/(0.8c)= 16 ns
before t=0. That corresponds to -9.6 nanoticks. As a result, the clock
at that point will then indeed indicate 0 - as it should be.

In the general case that both reference systems are in motion, all
measured lengths and durations will be distorted and not be conform
reality.
The Lorentz transformations (Poincare, 1905) follow directly from the
discussed model and were also used by Lorentz.

Harald van Lintel, Dec. 2004.

Martin Hogbin

unread,
Dec 19, 2004, 6:58:23 PM12/19/04
to

"Harry" <harald.v...@epfl.ch> wrote in message news:3bff5641.04121...@posting.google.com...

> In a recent posting I saw the claim:
>
> "Under SR, A thinks that B's ruler is shorter. B thinks that A's
> ruler is shorter.


> Under LET, A thinks that B's ruler is shorter. B thinks that A's
> ruler is *longer*."

That is wrong.

> "LET" referred to the mathematical physics as found in Lorentz' 1904
> paper, and "SR" referred to the mathematical physics of as found in
> Einstein's 1905 paper.
>

> After 1905 Lorentz taught relativity and he did
> not propose a separate "LET" as there is no difference as far as
> mathematical or experimental physics is concerned.

That is correct and well known.

Martin Hogbin

shuba

unread,
Dec 19, 2004, 7:01:06 PM12/19/04
to
Harald wrote:

> After 1905 Lorentz taught relativity and he did
> not propose a separate "LET" as there is no difference as far as
> mathematical or experimental physics is concerned.

Lorentz taught *general relativity*, which does not reduce to the
ether theory that you promote. That's a difference in the
physics. Take a look at Ilja Schmelzer's ideas, if you want to
see how an honest person conducts himself.


---Tim Shuba---

Androcles

unread,
Dec 19, 2004, 9:18:27 PM12/19/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:3bff5641.04121...@posting.google.com...
> In a recent posting I saw the claim:
>
> "Under SR, A thinks that B's ruler is shorter. B thinks that A's
> ruler is shorter.

That's because A and B are both stupid.

> Under LET, A thinks that B's ruler is shorter. B thinks that A's
> ruler is *longer*."

That's because Lorentz wants to explain MMX, which SR cannot.
However, anyone that thinks the Earth changes its shape as it rotates
doesn't know about geophysics and is stupid.

[snip - You won't learn]
Androcles.


Tom Roberts

unread,
Dec 19, 2004, 11:55:43 PM12/19/04
to
Harry wrote:
> In a recent posting I saw the claim:
> "Under SR, A thinks that B's ruler is shorter. B thinks that A's
> ruler is shorter.
> Under LET, A thinks that B's ruler is shorter. B thinks that A's
> ruler is *longer*."

This last claim is false. Do not believe everything you see in this
newsgroup.


> I don't think that that is correct; in any case, Lorentz obviously
> didn't feel that he should be stuck to some little errors and
> incomplete understanding that he had when writing a certain paper, and
> neither did Einstein. After 1905 Lorentz taught relativity and he did
> not propose a separate "LET" as there is no difference as far as
> mathematical or experimental physics is concerned.

Yes. Well known and non-controversial to anyone who has actually read
Lorentz's writings.


> That reciprocity of observations directly follows from Lorentz'length
> contraction and clock slowdown is not immediately obvious.

It is indeed obvious to anyone who understands group theory and the
structure of the Lorentz group. But I'll grant that to most people
around here that is by no means "immediate" (:-)).


Tom Roberts tjro...@lucent.com

Harry

unread,
Dec 20, 2004, 3:43:51 AM12/20/04
to

"Tom Roberts" <tjro...@lucent.com> wrote in message
news:jXsxd.2972$_X7....@newssvr33.news.prodigy.com...

Exactly. The reciprocity of "local time" observations wasn't clear to
Lorentz himself when he first presented it - as he admitted later. In
contrast, it was immediately obvious to Poincare when he read it (he even
illustrated the reciprocity that same year). But he was an excellent
mathematician, trained in group theory.

Harald


Harry

unread,
Dec 20, 2004, 3:55:46 AM12/20/04
to

"shuba" <tim....@lycos.ScPoAmM> wrote in message
news:tim.shuba-13CC7...@individual.net...

> ---Tim Shuba---

This thread is about the full observational reciprocity of "time dilation"
and "length contraction" independent of interpretation, and I don't know
enough of GRT to judge your claim. In any case, Lorentz obviously didn't
think so.

Harald


Harry

unread,
Dec 20, 2004, 4:05:36 AM12/20/04
to

"Androcles" <du...@dummy.net> wrote in message
news:TDqxd.559$ef...@fe1.news.blueyonder.co.uk...

Androcles, thanks for your great deduction that Lorentz was stupid.

Harald


Harry

unread,
Dec 20, 2004, 4:27:06 AM12/20/04
to

"Martin Hogbin" <goatN...@hogbin.org> wrote in message
news:cq54iv$kvs$1...@hercules.btinternet.com...

Yes. This thread is for the few who think that the measured effects can't
possibly be reciprocal.

Harald


Androcles

unread,
Dec 20, 2004, 7:14:10 AM12/20/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:41c694f6$1...@epflnews.epfl.ch...

You are welcome. Anytime you want some simple and straightforward
logical conclusions made, come and see me.
BTW, I think you mean "[snip - I won't learn]"
Androcles.

> Harald
>
>


kenseto

unread,
Dec 20, 2004, 4:01:08 PM12/20/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:3bff5641.04121...@posting.google.com...
> In a recent posting I saw the claim:
>
> "Under SR, A thinks that B's ruler is shorter. B thinks that A's
> ruler is shorter.
> Under LET, A thinks that B's ruler is shorter. B thinks that A's
> ruler is *longer*."

A correct ether theory would say that both A's ruler and B's ruler will
remain the same length physically. However, if A is more at rest in the
ether than B then the light path length of A's ruler is shorter than the
light path length of B's ruler.

>
> "LET" referred to the mathematical physics as found in Lorentz' 1904
> paper, and "SR" referred to the mathematical physics of as found in
> Einstein's 1905 paper.
>
> I don't think that that is correct; in any case, Lorentz obviously
> didn't feel that he should be stuck to some little errors and
> incomplete understanding that he had when writing a certain paper, and
> neither did Einstein. After 1905 Lorentz taught relativity and he did
> not propose a separate "LET" as there is no difference as far as
> mathematical or experimental physics is concerned.
>
> That reciprocity of observations directly follows from Lorentz'length
> contraction and clock slowdown is not immediately obvious.

There is no reciprocity in a correct ether theory. The rate of a clock is
dependent on the state of absolute motion of the clock and the clock at the
rest frame of the ether has the fastest clock rate. A correct ether theory
gives rise to an improved relativity theory called IRT. A description of IRT
is as follows:
Ken Seto
________________________________________________________________
IRT (Improved Relativity Theory) is a New Theory of Motion.
It includes SR and GR as subsets. Its equations are valid in all
environments....including gravity.
1) The laws of physics based on a clock second and light path
length of a measuring rod are the same for all observers in
all inertial reference frames.
2) The speed of light in free space based on a clock second
and the light path length of a measuring rod has the same
mathematical ratio c in all directions and all inertial frames.
3) The laws of physics based on a defined absolute second
and the physical length of a rod is different in different frames
of reference.
4) The one-way speed of light in free space based on a defined absolute
second and the physical length of a measuring rod has a different
mathematical ratio for light speed in different inertial frames. The
speed of light based on a defined absolute second and the physical
length of a measuring rod is maximum in the rest frame of the aether.

The Consequences of these Postulates:
(1). The speed of light is not a universal constant. It is a constant math
ratio as follows:
Light path length of rod (299,792,458 m)/the absolute time content for a
clock second co-moving with the rod.
Detailed explannation of this new definition:
By definition the speed of light in the rest frame of the ether is as
follows:
Light path length in the ether frame=gamma*299,792,458m. This is reduced
to--
(299,792,458m)-- because gamma is equal to one in the ether frame.
The absolute time content for a clock second in the ether frame=gamma*1
ether frame clock second. This is reduced to__ (1 ether frame clock
second)__ because gamma is equal to one in the ether frame.
Therefore the speed of light in the ether frame is:
299,792,458m/1 ether frame clock second
The speed of light in any frame moving in the stationary ether is determined
as follows:
The light path length of rod in the moving frame=gamma*299,792,458m
The absolute itme content for a moving clock second
=gamma*1 ether frame clock second
Therefore the speed of light in any moving frame in the stationary ether is
as follows:
gamma*299,792,458m/gamma*1 ether frame clock second.
This is reduce to a constant math ratio of:: 299,792,458m/1 ether frame
clock second

(2). The physical length of a rod remains the same in all frames of
reference. The light path length of a rod changes with the state of absolute
motion of the rod. The higher is the state of absolute motion the longer is
its light path length.
(3). The rate of a clock is dependent on the state of absolute motion of the
clock. The higher is the state of absolute motion the slower is its clock
rate.
(4). Absolute time exists. The relationship between clock time and absolute
time is as follows: A clock second will contain a different amount of
absolute time in different state of absolute motion (different frames of
reference). The higher is the state of absolute motion of the clock the
higher is the absolute time content for a clock second.
(5) Simultaneity is absolute. If two events are simultaneous in one frame,
identical events will also be simultaneous in different frames. However the
time interval for the simultaneity to occur will be different in different
frame. This is due to that different frames are in different states of
absolute motion.
(6) Relative motion between two observers A and B is the vector difference
of their absolute motions along the line joining A and B

C. The Math:

1. The time dilation (contraction) or expansion equations:
A and B are in relative motion from observer A's point of view:

Tab=Taa(Faa/Fab) OR Tab=Taa(Fab/Faa)

Taa=A clock time interval in observer A's frame
as measured by A
Tab= A's prediction of B's clock time interval for
an interval of Taa in his frame
Note: Even though Taa and Tab are two different clock time
intervals but in terms of absolute time content Taa=Tab

The light path length contraction or expansion equations for a physical rod:
Lab=Laa(Faa/Fab) OR Lab=Laa(Fab/Faa)

Laa=the light path length of a rod in A's frame as
measured by A.
Lab=the light path length of an identical rod in B's
frame as predicted by A
Note: Even though Laa and Lab are two different light
path lengths but these two light path lengths are
derived from identical rod that have the same
physical rod length. The different light path lengths
are the result of different states of absolute motion
of the rods.

2. The coordinate transform equations:
x'= Faa/Fab[x + t(Faa-Fab)(lambda)]
t'= Faa/Fab[t + x(Faa-Fab)/(Faa^2)(lambda)]
y'=y
z'=z

OR

x'= Fab/Faa[x - t(Faa-Fab)(lambda)]
t'= Fab/Faa[t - x(Faa-Fab)/(Faa^2)(lambda)]
y'=y
z'=z

A is the observer's frame (unprimed) and B is the observed frame (primed).
Faa = frequency of a standard light source in A's frame as measured by A.
Fab = frequency of an identical light source in B's frame as measured by A.
If Fab is not constant the mean value is used.
lambda = wave length of the standard light source in A's frame as measured
by A.

These coordinate transform equations are valid in all
environments ---including gravity. This means that
IRT includes SR/GR as subsets

3. Momentum of an object:
p=Mo(lambda)(Faa-Fab)

4. Kinetic Energy of an object::
K=Mo(Lambda)^2(Faa)^2(Faa/Fab-1)

5. Energy of a single particle:
E=Mo(Lambda)^2(Faa)^2

6. Gravtational Red or Blue Shift:
Delta (Faa) =Faa(1-Fab/Faa)
A positve value represents a red shift from A's location.
A negative value represents a blue shift from A's location

7. Gravitational Time Contraction or Expansion:
Delta(Taa)=Taa(1-Fab/Faa)
A positive value represents gravitational time contraction (dilation)
from A's location.
A negative value represents gravitational time expansion
from A's location.

8. The IRT procedure for determining the perihelion precession of
Mercury without recourse to GRT is:
a) Set up a corrdinate system for the Sun and Mervury.
b) Use the IRT Corrdinate tansformation equations to predict the
future positions of the Sun and Mercury.
c) The perihelion shift of Mercury will be revealed when these
future positions are plotted against time. Also, the value of the
shift can be determined from the plot.

Summarizing:
IRT is a complete theory of motion. It contains SR and GR as subsets.
It's equations are valid in all environments. In addition it resolves the
following observed difficulties of GR:
1. GR predicts that the expansion of the universe should be slowing down.
Actual observations show that the expansion is speeding up. IRT have no
such problem. It predicts what is observed.
2. GR gives the wrong prediction for the path of the space craft Pioneer 10.
IRT has no such problem. It predicts that the space craft is accelerating
toward the Sun because of a concentration of dark matter contained
within the solar system---especially around the Sun.
3. GR gives the wrong rotational curve for galaxies. Again IRT has no
such problem because IRT includes the effects of dark matter in its
calculations.

Ken Seto


harald.v...@epfl.ch

unread,
Dec 20, 2004, 5:30:25 PM12/20/04
to
Ken Seto wrote:
There is no reciprocity in a correct ether theory. The rate of a clock
is
dependent on the state of absolute motion of the clock and the clock at
the
rest frame of the ether has the fastest clock rate. A correct ether
theory
gives rise to an improved relativity theory called IRT. A description
of IRT
is as follows:

Ken, this thread is about how Lorentz contraction and clock slowdown
leads to reciprocity of observations when assuming an absolute
reference frame. I demonstrated it using a single reference system and
junior high-school math.

Harald

kenseto

unread,
Dec 20, 2004, 8:37:03 PM12/20/04
to

<harald.v...@epfl.ch> wrote in message
news:1103581825....@f14g2000cwb.googlegroups.com...

Sigh...your problem is that you assume v'=v. This assumption is what gives
you the reciprocity. In real life a clock second in the primed frame is
equal to
(Gamma*1 ether frame clock second). This means that v as measured by the
ether clock second is not equal to v' as measured by the moving frame clock
second.

Before you use high school math on me you need to learn some logic.

Ken Seto


kenseto

unread,
Dec 20, 2004, 9:09:53 PM12/20/04
to

<harald.v...@epfl.ch> wrote in message
news:1103581825....@f14g2000cwb.googlegroups.com...

another problem with your mathematical proof of reciprocity is that you
assumed that the primed frame (the moving frame) can be at rest and thus
have the same properties as the ether frame which logically sees all clocks
to run slow compared to its clock.

Ken Seto


Harry

unread,
Dec 21, 2004, 8:32:05 AM12/21/04
to

"kenseto" <ken...@erinet.com> wrote in message
news:37Lxd.4962$vb7...@fe1.columbus.rr.com...

>
> <harald.v...@epfl.ch> wrote in message
> news:1103581825....@f14g2000cwb.googlegroups.com...
> > Ken Seto wrote:
> > There is no reciprocity in a correct ether theory. The rate of a clock
> > is
> > dependent on the state of absolute motion of the clock and the clock at
> > the
> > rest frame of the ether has the fastest clock rate. A correct ether
> > theory
> > gives rise to an improved relativity theory called IRT. A description
> > of IRT
> > is as follows:
> >
> > Ken, this thread is about how Lorentz contraction and clock slowdown
> > leads to reciprocity of observations when assuming an absolute
> > reference frame. I demonstrated it using a single reference system and
> > junior high-school math.
>
> Sigh...your problem is that you assume v'=v.

Read it again. I did not assume but showed that v'=v.

Harry

unread,
Dec 21, 2004, 8:36:15 AM12/21/04
to

"kenseto" <ken...@erinet.com> wrote in message
news:RBLxd.3837$mA3...@fe2.columbus.rr.com...

>
> <harald.v...@epfl.ch> wrote in message
> news:1103581825....@f14g2000cwb.googlegroups.com...
> > Ken Seto wrote:
> > There is no reciprocity in a correct ether theory. The rate of a clock
> > is
> > dependent on the state of absolute motion of the clock and the clock at
> > the
> > rest frame of the ether has the fastest clock rate. A correct ether
> > theory
> > gives rise to an improved relativity theory called IRT. A description
> > of IRT
> > is as follows:
> >
> > Ken, this thread is about how Lorentz contraction and clock slowdown
> > leads to reciprocity of observations when assuming an absolute
> > reference frame. I demonstrated it using a single reference system and
> > junior high-school math.
>
> another problem with your mathematical proof of reciprocity is that you
> assumed that the primed frame (the moving frame) can be at rest

No the frame that is moving relative to the ether can never be at rest. I
showed that nevertheless it will appear to be at rest for co-moving
observers who assume to be at rest.

kenseto

unread,
Dec 21, 2004, 9:23:47 AM12/21/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:41c825e1$1...@epflnews.epfl.ch...

>
> "kenseto" <ken...@erinet.com> wrote in message
> news:RBLxd.3837$mA3...@fe2.columbus.rr.com...
> >
> > <harald.v...@epfl.ch> wrote in message
> > news:1103581825....@f14g2000cwb.googlegroups.com...
> > > Ken Seto wrote:
> > > There is no reciprocity in a correct ether theory. The rate of a clock
> > > is
> > > dependent on the state of absolute motion of the clock and the clock
at
> > > the
> > > rest frame of the ether has the fastest clock rate. A correct ether
> > > theory
> > > gives rise to an improved relativity theory called IRT. A description
> > > of IRT
> > > is as follows:
> > >
> > > Ken, this thread is about how Lorentz contraction and clock slowdown
> > > leads to reciprocity of observations when assuming an absolute
> > > reference frame. I demonstrated it using a single reference system and
> > > junior high-school math.
> >
> > another problem with your mathematical proof of reciprocity is that you
> > assumed that the primed frame (the moving frame) can be at rest
>
> No the frame that is moving relative to the ether can never be at rest.

Right.....when the moving observer sees the ether frame clock moving wrt him
it is not due to the ether frame clock that is moving but rather it is due
to the moving observer who is doing the moving. This means that the ether
frame clock can never be running slow compared to all the other clocks
moving in the ether. Also this means that there is no reciprocity between
the moving clock and the ether frame clock.

>I
> showed that nevertheless it will appear to be at rest for co-moving
> observers who assume to be at rest.

That's the problem. No observer can assume to be at rest in the ether. Also
There is no way to show that "it will appear to be at rest" without making
additional assumptions. Since no observer is at rest in the ether therefore
any observer will see some clocks moving wrt him running slow and some
clocks moving wrt him running fast.

Ken Seto


Harry

unread,
Dec 21, 2004, 9:59:11 AM12/21/04
to

"kenseto" <ken...@erinet.com> wrote in message
news:TlWxd.9744$LW1....@fe2.columbus.rr.com...

Right.

> This means that the ether
> frame clock can never be running slow compared to all the other clocks
> moving in the ether. Also this means that there is no reciprocity between
> the moving clock and the ether frame clock.

Right. If you had understood the text and calculations, you would now
*understand why* the moving measurement system will not only measure in
error, but that this error results in the misjudged assessment that the
ether frame clock is both moving and running slow, although this is
counter-intuitive as long as you don't grasp the timing effects.

> > I showed that nevertheless it will appear to be at rest for co-moving
> > observers who assume to be at rest.
>
> That's the problem. No observer can assume to be at rest in the ether.

Just ask around, and you will discover that nearly everyone thinks that
lightspeed is truly isotropic relative to the lab.

> Also
> There is no way to show that "it will appear to be at rest" without making
> additional assumptions. Since no observer is at rest in the ether
therefore
> any observer will see some clocks moving wrt him running slow and some
> clocks moving wrt him running fast.

Again you missed something. They have little other choice than to assume
that they are in rest when calibrating their system. Their only realistic
alternative would be to ask you how fast they are going! ;-)

Harald


kenseto

unread,
Dec 21, 2004, 9:45:27 AM12/21/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:41c824e7$1...@epflnews.epfl.ch...

>
> "kenseto" <ken...@erinet.com> wrote in message
> news:37Lxd.4962$vb7...@fe1.columbus.rr.com...
> >
> > <harald.v...@epfl.ch> wrote in message
> > news:1103581825....@f14g2000cwb.googlegroups.com...
> > > Ken Seto wrote:
> > > There is no reciprocity in a correct ether theory. The rate of a clock
> > > is
> > > dependent on the state of absolute motion of the clock and the clock
at
> > > the
> > > rest frame of the ether has the fastest clock rate. A correct ether
> > > theory
> > > gives rise to an improved relativity theory called IRT. A description
> > > of IRT
> > > is as follows:
> > >
> > > Ken, this thread is about how Lorentz contraction and clock slowdown
> > > leads to reciprocity of observations when assuming an absolute
> > > reference frame. I demonstrated it using a single reference system and
> > > junior high-school math.
> >
> > Sigh...your problem is that you assume v'=v.
>
> Read it again. I did not assume but showed that v'=v.

But what you showed is based some assumptions that leads to v'=v. In real
life the moving observer must assume that the observed relative motion
between him and the ether frame clock is due to his motion in the ether and
that the ether frame clock's motion is zero (v=0).

Ken Seto


Harry

unread,
Dec 21, 2004, 10:49:06 AM12/21/04
to

"kenseto" <ken...@erinet.com> wrote in message
news:bGWxd.5932$vb7....@fe1.columbus.rr.com...

If we do that, we will actually be using the ether as master reference
frame. And with such a reference, everyone will agree. Now tell me your
speed relative to the ether (we need to know it accurately - we want at
least 4 decimals precision).

Harald


kenseto

unread,
Dec 21, 2004, 1:17:03 PM12/21/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:41c84503$1...@epflnews.epfl.ch...

There is no need to know your speed in the ether. You can do calculations
with IRT and give you all the answers you needed. However, there is a way to
determine your absolute motion. The experiment described in the following
link (page 3) is such an experiment.

http://www.journaloftheoretics.com/Links/Papers/Seto.pdf

Ken Seto


kenseto

unread,
Dec 21, 2004, 2:35:08 PM12/21/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:41c83950$1...@epflnews.epfl.ch...

Sigh....there is no measurement make between frames. There are predictions
by the various theories. The calculations you presented is based on the
euqations which are in turn based on a set of assumptions that already
assumed reciprocity.

>but that this error results in the misjudged assessment that the
> ether frame clock is both moving and running slow, although this is
> counter-intuitive as long as you don't grasp the timing effects.

A correct ether such as IRT would avoid such misjudged assessment. IRT has
the correct assement of the situation. It says that a moving observer will
see the ether frame clock not to be moving and the clock rate at the ether
frame is faster than his clock.


>
> > > I showed that nevertheless it will appear to be at rest for co-moving
> > > observers who assume to be at rest.
> >
> > That's the problem. No observer can assume to be at rest in the ether.
>
> Just ask around, and you will discover that nearly everyone thinks that
> lightspeed is truly isotropic relative to the lab.

So?? A correct ether theory would predict that the speed of light to be
isotropic in all frames. The reason is that the speed of light is a constant
math ratio as follows:
Light path length of rod (299,792,458m)/the absolute time content for a


clock second co-moving with the rod.
>

> > Also
> > There is no way to show that "it will appear to be at rest" without
making
> > additional assumptions. Since no observer is at rest in the ether
> therefore
> > any observer will see some clocks moving wrt him running slow and some
> > clocks moving wrt him running fast.
>
> Again you missed something. They have little other choice than to assume
> that they are in rest when calibrating their system.

I didn't miss anything. IRT allows the observer's clock to run fast or slow
compared to a clock moving wrt him.

>Their only realistic
> alternative would be to ask you how fast they are going! ;-)

NO...IRT is a realistic alternative. You can determine how fast you are
moving in the ether by doing the experiment in the following link (page 3).

Ken Seto


Harry

unread,
Dec 22, 2004, 9:16:23 AM12/22/04
to

"kenseto" <ken...@erinet.com> wrote in message
news:MV_xd.11885$mA3....@fe2.columbus.rr.com...

>
> "Harry" <harald.v...@epfl.ch> wrote in message
> news:41c83950$1...@epflnews.epfl.ch...
> >
> > "kenseto" <ken...@erinet.com> wrote in message
> > news:TlWxd.9744$LW1....@fe2.columbus.rr.com...
SNIP

> > > This means that the ether
> > > frame clock can never be running slow compared to all the other clocks
> > > moving in the ether. Also this means that there is no reciprocity
> between
> > > the moving clock and the ether frame clock.
> >
> > Right. If you had understood the text and calculations, you would now
> > *understand why* the moving measurement system will not only measure in
> > error,
>
> Sigh....there is no measurement make between frames. There are predictions
> by the various theories. The calculations you presented is based on the
> euqations which are in turn based on a set of assumptions that already
> assumed reciprocity.

Some here emphatically deny the reciprocity following from the proposed
effects. That's why I gave the example.

> >but that this error results in the misjudged assessment that the
> > ether frame clock is both moving and running slow, although this is
> > counter-intuitive as long as you don't grasp the timing effects.
>
> A correct ether such as IRT would avoid such misjudged assessment.

Fabulous. I'm amazed about your God who is so thoughtful to construct an
ether that will avoid misjudged assessments of it.

> IRT has
> the correct assement of the situation. It says that a moving observer will
> see the ether frame clock not to be moving and the clock rate at the ether
> frame is faster than his clock.

Then you're a wonderboy to think that you can tell from your observation
which things are in rest in the ether - then we can base our reference
system on that!

> > > > I showed that nevertheless it will appear to be at rest for
co-moving
> > > > observers who assume to be at rest.
> > >
> > > That's the problem. No observer can assume to be at rest in the ether.
> >
> > Just ask around, and you will discover that nearly everyone thinks that
> > lightspeed is truly isotropic relative to the lab.
>
> So??

So your "no observer" is wrong, there are many observers who make an
assumption that is effectively the same.

> A correct ether theory would predict that the speed of light to be
> isotropic in all frames.

An example of a visible ether is air. Now give a "correct" theory for sound
propagation in which the speed of sound around a riding train is isotropic
relative to the train....

> The reason is that the speed of light is a constant math ratio as follows:
> Light path length of rod (299,792,458m)/the absolute time content for a
> clock second co-moving with the rod.

Physics isn't math. But then, many in this group don't understand that...

> > > Also
> > > There is no way to show that "it will appear to be at rest" without
> making
> > > additional assumptions. Since no observer is at rest in the ether
> > therefore
> > > any observer will see some clocks moving wrt him running slow and some
> > > clocks moving wrt him running fast.
> >
> > Again you missed something. They have little other choice than to assume
> > that they are in rest when calibrating their system.
>
> I didn't miss anything. IRT allows the observer's clock to run fast or
slow
> compared to a clock moving wrt him.
>
> >Their only realistic
> > alternative would be to ask you how fast they are going! ;-)
>
> NO...IRT is a realistic alternative. You can determine how fast you are
> moving in the ether by doing the experiment in the following link (page
3).

I have seen it a long time ago, but I gave up trying to explain it to you.

All the best,
Harald


kenseto

unread,
Dec 22, 2004, 2:50:37 PM12/22/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:41c980c7$1...@epflnews.epfl.ch...

>
> "kenseto" <ken...@erinet.com> wrote in message
> news:MV_xd.11885$mA3....@fe2.columbus.rr.com...
> >
> > "Harry" <harald.v...@epfl.ch> wrote in message
> > news:41c83950$1...@epflnews.epfl.ch...
> > >
> > > "kenseto" <ken...@erinet.com> wrote in message
> > > news:TlWxd.9744$LW1....@fe2.columbus.rr.com...
> SNIP
>
> > > > This means that the ether
> > > > frame clock can never be running slow compared to all the other
clocks
> > > > moving in the ether. Also this means that there is no reciprocity
> > between
> > > > the moving clock and the ether frame clock.
> > >
> > > Right. If you had understood the text and calculations, you would now
> > > *understand why* the moving measurement system will not only measure
in
> > > error,
> >
> > Sigh....there is no measurement make between frames. There are
predictions
> > by the various theories. The calculations you presented is based on the
> > euqations which are in turn based on a set of assumptions that already
> > assumed reciprocity.
>
> Some here emphatically deny the reciprocity following from the proposed
> effects. That's why I gave the example.

When both observers use the same equation to calculate clock time in each
other frame that means that you already assumed reciprocity


>
> > >but that this error results in the misjudged assessment that the
> > > ether frame clock is both moving and running slow, although this is
> > > counter-intuitive as long as you don't grasp the timing effects.
> >
> > A correct ether such as IRT would avoid such misjudged assessment.
>
> Fabulous. I'm amazed about your God who is so thoughtful to construct an
> ether that will avoid misjudged assessments of it.

IRT realizes that an observed clock can run fast or slow compared to the
observer's clock but it does not say which clock run slow and that's why it
gives two equations for clock time dilation.


>
> > IRT has
> > the correct assement of the situation. It says that a moving observer
will
> > see the ether frame clock not to be moving and the clock rate at the
ether
> > frame is faster than his clock.
>
> Then you're a wonderboy to think that you can tell from your observation
> which things are in rest in the ether - then we can base our reference
> system on that!

All objects are in a state of absolute motion. No object is at rest in the
ether. That's why IRT give two equations....one for clock time dilation and
the other for clock time expansion.
>

> > > Just ask around, and you will discover that nearly everyone thinks
that
> > > lightspeed is truly isotropic relative to the lab.
> >
> > So??
>
> So your "no observer" is wrong, there are many observers who make an
> assumption that is effectively the same.

Where did I say "no observer"??


>
> > A correct ether theory would predict that the speed of light to be
> > isotropic in all frames.
>
> An example of a visible ether is air.

That's not the correct structure for an ether.

> Now give a "correct" theory for sound
> propagation in which the speed of sound around a riding train is isotropic
> relative to the train....

The speed of light is isotropic on earth because the direction of absolute
motion is in the vertical direction.


>
> > The reason is that the speed of light is a constant math ratio as
follows:
> > Light path length of rod (299,792,458m)/the absolute time content for a
> > clock second co-moving with the rod.
>
> Physics isn't math. But then, many in this group don't understand that...

The above is not math. It is a definition.


>
> > > Again you missed something. They have little other choice than to
assume
> > > that they are in rest when calibrating their system.
> >
> > I didn't miss anything. IRT allows the observer's clock to run fast or
> slow
> > compared to a clock moving wrt him.
> >
> > >Their only realistic
> > > alternative would be to ask you how fast they are going! ;-)
> >
> > NO...IRT is a realistic alternative. You can determine how fast you are
> > moving in the ether by doing the experiment in the following link (page
> 3).
>
> I have seen it a long time ago, but I gave up trying to explain it to
you.

I don't recall that. I doubt that you have read it. In any case it is
approved for publication by the peer reviewed journal "Galilean
Electrodynamics". I value their judgement more than your ranting.

Ken Seto


Harry

unread,
Dec 23, 2004, 4:43:57 AM12/23/04
to

"kenseto" <ken...@erinet.com> wrote in message
news:hekyd.20626$mA3....@fe2.columbus.rr.com...

>
> "Harry" <harald.v...@epfl.ch> wrote in message
> news:41c980c7$1...@epflnews.epfl.ch...

SNIP because I give up again. ONly one remark:

> > I have seen it a long time ago, but I gave up trying to explain it to
> you.
>
> I don't recall that. I doubt that you have read it. In any case it is
> approved for publication by the peer reviewed journal "Galilean
> Electrodynamics". I value their judgement more than your ranting.

AFAIK that journal didn't peer review it and then "approved it for
publication" .

Harald


kenseto

unread,
Dec 23, 2004, 9:44:20 AM12/23/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:41ca...@epflnews.epfl.ch...

Idiot.

Ken Seto

Harry

unread,
Dec 23, 2004, 12:27:20 PM12/23/04
to

"kenseto" <ken...@erinet.com> wrote in message
news:8RAyd.21340$mA3....@fe2.columbus.rr.com...

* Not peer-reviewed or edited, but thought to be worthwhile and interesting
enough for easy public access. These articles are not copyrighted and
remain the author's property.
- http://www.journaloftheoretics.com/Links/links-papers.htm

I wonder if he carefully read it; in any case your statement was misleading.

Harald


kenseto

unread,
Dec 23, 2004, 1:33:43 PM12/23/04
to

"Harry" <harald.v...@epfl.ch> wrote in message
news:41caff05$1...@epflnews.epfl.ch...

Hey idiot....that's not the Peer-Rewieved journal "Galilean
Electrodynamics". They plan to public my experiment in 2005.

Ken Seto


harry

unread,
Dec 28, 2004, 6:48:05 AM12/28/04
to
Here follows an addition to my example, in which I show that no light
assumption is needed - specially written for Greywolf. :-))

Another, less used way of synchronizing clocks is "slow" clock
transport. We will verify if it should yield the same result as the
"esync" in our example.

A co-moving observer calibrates a clock C3 directly next to clock C1
and then moves it to the rear clock C2 at 0.001c according to speed
measurement in that system. We supposing that the system has been
calibrated before by esync.

Then, as the co-moving observer assumes to be in rest, he will expect
that it will take 6 marks / 3E5 = 20 microticks.
C3 is supposed to advance SQR(1 - (0.001)^2) * 20 ticks =
= (20 microticks - 10 pticks). Thus C2 should indicate 10 pticks more
than C3.

In reality the clock transport will take (20 - 0.016)= 19.084
microticks on C1 and C2, corresponding to 33306.7 ns.
The ruler travels in this time 33.3067E-6 * 0.8c = 7993.6 m, so that C3
travels from C1 by (7993.6 - 3.6) m = 7990 m through the ether to C2.

C3 will thus advance by:
SQRT(1 - 0.7996)^2) * 33306.7 ns = 19855.990 nanoticks.
C1 advanced by: 19840.000 nanoticks
...................................--------- -
Advance of C3 on C1: 15.990 nanoticks
Advance of C2 : 16.000 nanoticks
......................................------ -
Advance of C3 on C2 : -10 picoticks

Thus if C2 has an offset of 16 nanoticks, C3 will lag 10 pticks behind
on C2 and the moving observer will consider C2 to be perfectly
synchronized.

Harald, Dec. 2004

greywolf42

unread,
Dec 28, 2004, 1:55:40 PM12/28/04
to
harry <harald.v...@epfl.ch> wrote in message
news:1104234485.6...@z14g2000cwz.googlegroups.com...

> Here follows an addition to my example, in which I show that no light
> assumption is needed - specially written for Greywolf. :-))
>
> Another, less used way of synchronizing clocks is "slow" clock
> transport. We will verify if it should yield the same result as the
> "esync" in our example.

Slow clock transport is not a way of synchronizing clocks. It is supposed
to be a way of maintaining e-synchronization.

> A co-moving observer calibrates a clock C3 directly next to clock C1

You mean sets it to the same time?

> and then moves it to the rear clock C2 at 0.001c according to speed
> measurement in that system.

Sorry, can't do it. We haven't measured 'c' yet. That's the whole point of
the experiment.

> We supposing that the system has been
> calibrated before by esync.

Yep. That's where you make your error. Nothing further must be said. Once
you assume e-synching, you are using light to measure distances. And you'll
get exactly what Dr. E claimed. Not because the second postulate is
correct, but because e-synching forces any readings to match.

{snip the rest}

--
greywolf42
ubi dubium ibi libertas
{remove planet for return e-mail}

harry

unread,
Dec 28, 2004, 5:20:42 PM12/28/04
to

greywolf42 wrote:
> harry <harald.v...@epfl.ch> wrote in message
> news:1104234485.6...@z14g2000cwz.googlegroups.com...
> > Here follows an addition to my example, in which I show that no
light
> > assumption is needed - specially written for Greywolf. :-))
> >
> > Another, less used way of synchronizing clocks is "slow" clock
> > transport. We will verify if it should yield the same result as the
> > "esync" in our example.
>
> Slow clock transport is not a way of synchronizing clocks. It is
supposed
> to be a way of maintaining e-synchronization.

Nonsense. No reference to e-synching is needed.

> > A co-moving observer calibrates a clock C3 directly next to clock
C1
>
> You mean sets it to the same time?

Yes. DOn't you know how slow clock transport works?!

> > and then moves it to the rear clock C2 at 0.001c according to speed
> > measurement in that system.
>
> Sorry, can't do it. We haven't measured 'c' yet. That's the whole
point of
> the experiment.

I showed in the foregoing example that c'= c. But here it doesn't
really matter. Just move the clock slow and next time slower to see if
it was slow enough. The slowest movement will show that the motion of
the example has a 10 ps deviation due to "fast" motion relative to the
ruler. All this is well known and you can verify it in one minute...

> > We supposing that the system has been
> > calibrated before by esync.
>
> Yep. That's where you make your error. Nothing further must be
said. Once
> you assume e-synching, you are using light to measure distances.

Of course not, no e-synching is needed for this method. I just showed
that you are mistaken to claim that the result will *differ*.

> And you'll
> get exactly what Dr. E claimed. Not because the second postulate is
> correct, but because e-synching forces any readings to match.

Dr. E and e-synching have nothing to do with this. But your
presumptions and refusal to pick up a pocket calculator have everything
to do with it...
As long as you don't verify these things for yourself you are beyond
help. I thought that by showing you step by step how it works you would
get it, but I used bad psychology by using the "esync" word
unnecessarily. Sorry. Thus I wasted my time on you, but no hard
feelings. Next time you argue the same old nonsense about
non-symmetrical measurements I'll simply ignore it.

Cheers,
Harald

greywolf42

unread,
Dec 30, 2004, 1:15:05 PM12/30/04
to
harry <harald.v...@epfl.ch> wrote in message
news:1104272442.6...@f14g2000cwb.googlegroups.com...

>
> greywolf42 wrote:
> > harry <harald.v...@epfl.ch> wrote in message
> > news:1104234485.6...@z14g2000cwz.googlegroups.com...

> > > Here follows an addition to my example, in which I show that no
> > > light assumption is needed - specially written for Greywolf. :-))
> > >
> > > Another, less used way of synchronizing clocks is "slow" clock
> > > transport. We will verify if it should yield the same result as the
> > > "esync" in our example.
> >
> > Slow clock transport is not a way of synchronizing clocks. It is
> > supposed to be a way of maintaining e-synchronization.
>
> Nonsense. No reference to e-synching is needed.

SCT *is* used with e-synching. It is not itself a way of synchronizing
clocks. And you admitted assuming that the clocks had been e-synched,
below.

> > > A co-moving observer calibrates a clock C3 directly next to clock
> > > C1
> >
> > You mean sets it to the same time?
>
> Yes. DOn't you know how slow clock transport works?!

I was merely attempting to clarify your unusual term of 'calibration' for
'synchronization'. The two terms mean substantially different things.

> > > and then moves it to the rear clock C2 at 0.001c according to speed
> > > measurement in that system.
> >
> > Sorry, can't do it. We haven't measured 'c' yet. That's the whole
> > point of the experiment.
>
> I showed in the foregoing example that c'= c.

There *IS* no foregoing example in this process. c' = c only if you perform
e-synching.

> But here it doesn't
> really matter. Just move the clock slow and next time slower to see if
> it was slow enough. The slowest movement will show that the motion of
> the example has a 10 ps deviation due to "fast" motion relative to the
> ruler. All this is well known and you can verify it in one minute...

How can it be verified? It has never been done without e-synching.

(A)


> > > We supposing that the system has been
> > > calibrated before by esync.
> >
> > Yep. That's where you make your error. Nothing further must be
> > said. Once you assume e-synching, you are using light to measure
distances.

(B)


> Of course not, no e-synching is needed for this method. I just showed
> that you are mistaken to claim that the result will *differ*.

Your statement (A) directly contradicts your statement (B).

> > And you'll
> > get exactly what Dr. E claimed. Not because the second postulate is
> > correct, but because e-synching forces any readings to match.
>
> Dr. E and e-synching have nothing to do with this. But your
> presumptions and refusal to pick up a pocket calculator have everything
> to do with it...

> As long as you don't verify these things for yourself you are beyond
> help. I thought that by showing you step by step how it works you would
> get it, but I used bad psychology by using the "esync" word
> unnecessarily. Sorry. Thus I wasted my time on you, but no hard
> feelings. Next time you argue the same old nonsense about
> non-symmetrical measurements I'll simply ignore it.

At least you'll be consistent in ignoring everything you don't want to see.
;)

harry

unread,
Dec 31, 2004, 11:28:19 AM12/31/04
to
Harald wrote:
> But here it doesn't
> really matter. Just move the clock slow and next time slower to see
if
> it was slow enough. The slowest movement will show that the motion of
> the example has a 10 ps deviation due to "fast" motion relative to
the
> ruler. All this is well known and you can verify it in one minute...

Greywolf:


How can it be verified? It has never been done without e-synching.
(A)

Hey I here above *explained* how it is done (if you take the effort to
grab a calculator...).
Now I guess that you really don't know slow clock transport, despite my
explanation of how it works.
Please allow me a last attempt to activate your grey cells to switch
from auto-pilot into think-mode:
According to you, if you slowly move a clock through the ether from one
clock to another, what is the problem, in principle? How bad would it
be in practice? And how can you determine how bad it is without knowing
c?
(I already gave you the answers, but discovered that it's useless as
long as you don't find out for yourself!).

> > > We supposing that the system has been
> > > calibrated before by esync.

> > Yep. That's where you make your error. Nothing further must be
> > said. Once you assume e-synching, you are using light to measure
distances
.
(B)
> Of course not, no e-synching is needed for this method. I just showed
> that you are mistaken to claim that the result will *differ*.

Greywolf wrote:
Your statement (A) directly contradicts your statement (B).

In other words:

(1) you can, if you are at rest in the ether, synchronize clocks with
esynching.

(A) you can also, if you are at rest in the ether, synchronize clocks
without esynching and use slow clock transport instead - the slower the
more accurate, and how slow that should be is easy to discover by
trying.

(B) If you transport the clock very slowly, you will find no difference
between the two methods (1) and (A).

Where's the contradiction? And how can it be that you don't know this
after all these years?!

> > And you'll
> > get exactly what Dr. E claimed. Not because the second postulate is
> > correct, but because e-synching forces any readings to match.

> Dr. E and e-synching have nothing to do with this. But your
> presumptions and refusal to pick up a pocket calculator have
everything
> to do with it...
> As long as you don't verify these things for yourself you are beyond
> help. I thought that by showing you step by step how it works you
would
> get it, but I used bad psychology by using the "esync" word
> unnecessarily. Sorry. Thus I wasted my time on you, but no hard
> feelings. Next time you argue the same old nonsense about
> non-symmetrical measurements I'll simply ignore it.

Greywolf:


At least you'll be consistent in ignoring everything you don't want to
see.
;)

I wasted my time to make *you* see what you obviously *refused* to see
- so much that you even couldn't get yourself to plug the numbers in...
likely you're too afraid that it may be correct. It must be
psychological!
There were a few typos in the calculation but you didn't notice (it
should have been 19984 and 19999.990 nanoticks for C1 and C3
respectively) . And will you even *try* to check it? I don't think so!
But just in case, in order to know what I'm talking about, you only
have to verify that if C3 travels in 33307 nanoseconds through the
ether from C1 to C2, in that time C1 will advance by 19984 nanoticks
but C3 by 20000 nanoticks. And that the error would be 0 nanoticks if
the ruler were resting in the ether.

Harald

Reply all
Reply to author
Forward
0 new messages