Wednesday, March 16, 2005

Evidence for "Accelerating Universe" Questionable

Theoretists must be careful not to go too far, before coming back to re-exam some of the most crucial experimental evidences presented and see exactly how trust worthy those data are.

The original paper presenting the "most reliable" evidence of an "accelerating universe" is none other than this, if you have the patience to go over 53 pages of technicalities of data filtration and manipulation tricks to extract the useful data:

I must point out the fact that NOT A SINGLE paper has published questioning the validity of the data or the method used in extracting the data. Such a unanimousity of voice regarding a piece of crucial experimental data is dangerous, especially consider how miniscure the data is.

Put it this way: The whole "accelerating universe" idea that the whole science community is talking about, is based on no more than a few dozen stray photons collected over the course of a couple hundred hours exposure time, and extracted from amoung other heavenly photons, using very complicated computation models.

Let's look at two images. First this one is the 1995 image, before the allerged supernova:

This is the 1997 image, where the supernova happened:

Compare the two, you can identify the SAME insignificantly dim yellow dots, seemingly have almost the same brightness, before and after the supernova. (if you have difficulty identifying the dot in the first image, just start from the center and search along the 3 o'clock or 3:30 direction, you will see it).

That tiny little dot is not even the supernova itself. It is merely the remote galaxy the supernova sits in. That dot is almost of the same brightness, meaning that photons attributable to the galaxy ifself far exceeds photons attributable to the supernova.

The whole galaxy is only one dot of no more than a few pixels. How would you extract photos from that supernova, and distinguish them from the background of the galaxy, which is one or two orders more photons. And be able to analysis the spectrum of the supernova, and tell what type it is?

The author claimed the data would require a resolution of better than 1/10 of one pixel, to be able to distinguish SN photons from the (much stronger) background (of the host galaxy). That's a stretch 10 times better than the optical resolution itself, and requires impossible stability of the Hubble to not have shaked a bit during several months of observation.

The photon counts is just too low to be much meaningful, IMO. During hundreds of hours of observation, one stray photons bunced by a space dust is enough to leave a dot on the image.

Why no one has ever raised the question of the credibility of the data?