WRONG PARADIGM BLUES
When you are working within the wrong conceptual paradigm things do not work very well. Observations tend to be embarrassingly at odds with theoretical expectations. Predicted phenomena are AWOL. Epicycles have to be added on to your models until they become ungainly and downright ugly.
For the last century or so physics has totally bought into the idea of strict reductionism, which means that things always get more fundamental as you go to smaller scales. There has been some interest in emergent modelling, but this still puts the most fundamental phenomena on the smallest scales.
In a discrete global fractal paradigm, reductionism works very well within the limited context of single cosmological Scales, such as the Atomic Scale, Stellar Scale or Galactic Scale. But crucially, this reductionism does not work between and among cosmological Scales. Each cosmological Scale is equally fundamental, i.e., none is more fundamental than any other.
This is a radically new way of looking at nature, i.e., a new over-arching cosmological paradigm. Dimensional constants (except velocities like c which are completely scale invariant) are different on different Scales. The gravitational constant, for example, has values that differ by ~10³⁸ on neighboring Scales. This changes everything in a profound manner. It yields a far more natural, physical and sensible Planck mass and length. It radically changes the dynamics of the microcosm.
A huge amount of empirical and theoretical evidence supporting the discrete fractal paradigm can be found at http://www3.amherst.edu/~rloldershaw . Any question one might ask is answered there in one or more places. I am also more than willing to help with explanations or listen to thoughtful comments and criticism.
We desperately need a new paradigm in theoretical physics. Not a new mathematical model detached from observations, not a highly abstract theory that takes 10 years to master, but a new and comprehensible conceptual paradigm that is founded on observational knowledge, that is far more unified than anything we now have, that makes testable predictions, and that actually successfully passes those predictive tests.