I have read a little bit about the structural realist's position. It is characterised by the belief that there are key ideas that must be retained in scientific theories, namely those mathematical/theoretical 'structures' which made a theory empirically successful since new theories must maintain and build upon the empirical success of older theories.
However, I'm having a hard time interpreting this idea of the retention of a 'mathematical structure'. In Worrall's paper, he uses the example of light as we transitioned from a physical ether theory to Maxwell's theory of electromagnetism. In this transition the 'entities' in the theory changed. In the ether theory, light was taken to be oscillations of some physical medium and was taken my Maxwell to be the oscillations in some field (which seems to be a much more abstract and theoretical concept). In either case, light is taken to be this kind of oscillating thing.
Could someone elaborate on this idea? Are there other examples from theories of physics of this kind of thing?
What about cases where theories have been overtaken by new theories from which the principles of the old theory can be derived?
E.g. You can quite trivially derive Kepler's 3rd law of planetary motion from Newtonian Gravity and motion (which was the 'replacement' for Kepler's laws more or less). Is this an instance of some kind of retention of mathematical structure or am I misunderstanding things?