On 04/19/2010 09:53 PM, Harry Veeder wrote:
> > His interest is now in the movement of the fringes rather than in the > number of fringes: > > "The zero point, where a standstill of the pattern movement happens, > is for both interferometers at the same position. There are two zero > points in a 360° rotation of both interferometers when the beam > splitter is positioned horizontally to the earth's surface. > > To all people who say that the only influence on the interferometer > is gravity, Well, really, I mean there's been a gross error detected by skeptics in the first design, which he proved was indeed gross by reducing it by a lot with a stiffer apparatus. So, obviously, if he's going to claim the residual shift isn't also due to flexing, he'd better have some means of measuring the stiffness of the apparatus, or some way of predicting the amount of flex from first principles. Without *something* to address flexing, it's hard to take this too seriously. In other words, he'd better figure out where his error bars are if he wants anybody to take this seriously. An analysis which goes "It's been redesigned and now it's real stiff, for sure, so we can ignore flexing" isn't very convincing. That's tantamount to claiming his error bars are *zero* size. > I have a simple question. Why is there no zero point or > stop of the fringe pattern movement when the beam splitter is in the > vertical position? In the beam splitter's vertical position, the > mirrors and the mirror holders are symmetrically pushed by gravity. > But there is no zero point." I'm not sure I understand this question. Of course there are going to be two extrema in the shift, which correspond to "zero points" in the fringe motion. The fact that the extrema occur 180 degrees apart doesn't seem surprising. Determining *where* we'd expect them to occur would require a careful and detailed mechanical analysis of the apparatus, and might require some careful measurements of flexing of the components. He seems to be speculating that there should be *four* extrema as the apparatus is rotated. Before we get too surprised that there are just two extrema, one would like to see somewhat more solid reasoning than "the mirrors and holders are symmetrically pushed by gravity". It's cool that he's putting in the time and effort to do this, but there seems to be a basic in his whole approach. He's taking the view that any apparently anomalous result is *probably* due to a deep flaw in conventional physics theory. In other words, he's apparently making no effort to criticize his results himself -- he's doing no self-checking, and making no effort to estimate the expected errors in his measurements. This is not reasonable. He should start by questioning everything about the experiment himself, try to estimate the errors himself, and then, when he's have proved to his own satisfaction that his results differ from theory by MORE THAN THE SIZE OF HIS ERROR BARS, announce it. The approach he's taking seems to be very much like Naudin's approach, which always yields a positive result -- all Naudin's experiments seem to show OU behavior of something. FWIW scientists trying to prove conventional theory *right* are occasionally guilty of the same sort of behavior. The famous H-K experiment done with flying clocks seems to be an example, where they should have deconstructed their own experiment before concluding that it proved something. In fact their error bars may have been so large that the 'null hypothesis' fit their data just as well as SR's prediction, but since they didn't initially publish either the raw data or a careful error analysis it was hard to tell.
