There's probably some formal definition from Information Theory, but I'm too lazy to look it up. My working definition is just the number of bits it takes to accurately model a system.

You're right about science lending itself to false ideas and dogmas, but you have to keep in mind that science should always be a response to observation and experiment. Scientific theories are formulated to explain the data, and they are valuable only as long as they can be used to suggest ways they can themselves be tested. Modern ideas like string theory are ways to explain what we see, but they should never be mistaken for explanations. Sometimes testing a theory is not possible, we don't have the technology required to set up the appropriate experiment, or our telescopes and particle accellerators aren't up to the task of verifying some prediction the theory makes. Until a theory can be tested, we should never mistake it for reality. Even after it is tested successfully, we have no guarantee that future tests will not show areas where it doesn't work and has to be modified or discarded.

To pick an example from safely in the past, A geocentric universe as believed in pretty generally until Copernicus was a quite useful theory. It allowed us to navigate ships and predict eclipses, keep time and construct calenders. In order to meet the demands of more and better astronomical data, it was necessary to modify and amend this theory so much that soon it became obvious that there was a serious problem with it. Eventually, a paradigm shift was necessary to accomodate new theories which could do the explaining better, but even many of these were not as accurate as the old, "incorrect" theory was at predicting some phenomena. Copernicus gave us a heliocentric universe, Kepler gave us a mathematical description for it that worked, Newton showed us a physical mechanism to explain WHY the geometrical features manifested themselves the way they did, and even Newtonian mechanics was replaced by Special and the General Relativity. We have no guarantee the process has ended there, or that it will ever end, or even if it is fundamentally possible for it to end. Indeed, since the human mind is a subset of the cosmos, it is not unreasonable to assume that it will never be able to understand the universe at a deep and complete level.

The symbiosis between theoretical and experimental/observational science is essential to progress, one cannot exist without the other, and as long as the latter is around to test the former, there will be progress. The experimentalist can shoot down a theory, or partially confirm it, but the theoretican has to try and provide the inspiration for the observer to seek out the verifying data. Today, as you point out, there are a lot of theories floating around; they are all clever and creative and some sound very convincing, but they are really nothing but speculation and we should never forget that.

That's why sometime they sounds like false ideas and dogma: in a way they are. Edwin Hubble's 1936 book "The Realm of the Nebulae", ends with the following quote, which pretty well sums it all up and precisely describes the situation we find ourselves in today.

"Thus the explorations of space end on a note of uncertainty. And necessarily so. We are, by definition, in the very center of the observable region. We know our immediate neighborhood rather intimately. With increasing distance, our knowledge fades, and fades rapidly. Eventually, we reach the dim boundary--the utmost limits of our telescopes. There, we measure shadows, and we search among ghostly errors of measurement for landmarks that are scarcely more substantial.

"The search will continue, not until the empirical resources are exhausted, need we pass on to the dreamy realms of speculation."