Writing helps me think, so I write these notes more for me than for you. Do feel free to comment if you'd like.
A theory is simply a proposition that explains one or more observed phenomena. What makes a theory useful is that it can predict future observations. What makes a theory scientific is that these predictions can be confirmed or denied.
A theory is a scientific theory if it can be proven wrong.
That is a simple requirement and it is the only requirement for a theory to be a scientific theory.
Creationism is a theory because it is used to explain the existence of the Universe. It is not a scientific theory though because it can not be proven wrong.
Darwin's original theory of the origin of the species is a scientific theory and has been proven to be wrong in many respects. It has been modified thanks to the advent of genetic theory and the understanding of mutations that can be passively passed from generation to generation only making themselves evident in rare occasions.
Newton's theory of gravity has been proven wrong. It has been modified thanks to our understanding of the restrictions that relativity has put on it and that led us to the Einstein's theory of general relativity.
In turn Einstein's theory of general relativity has been proven wrong. It breaks down at the very small level where quantum theory takes over.
Quantum theory has also been proven wrong. It breaks down at large scales where general relativity takes over. At present we do not have a viable theory of quantum gravity. There are many candidates out there but we do not have enough data to differentiate between them.
Does all this mean that science is worthless? After all, there isn't a single scientific theory that has not been proven wrong.
No, the pursuit of science is not worthless. We have the theories and we understand in what areas those theories are valid. We use Newtonian gravity to calculate the orbits of the planets, moons and man made satellites... but we know well that if we want those calculations to be extremely accurate (such us for GPS satellites) then we need to include relativistic effects.
There is probably no one reading this far along my ramblings, but I am writing mostly for myself and I find it always a good idea to summarize:
To make a theory useful it must make predictions. To make a theory scientific those predictions must be falsifiable. A scientific theory can never be proven right. If a theory's prediction is wrong then the theory is wrong. If a theory's prediction is correct, then we make another prediction and try to prove it wrong again.
My next note will be about what exactly happens in practice, because as I mentioned, every theory I know of has been proven wrong... but we still use them and successfully so.
(Cross-posted by permission of the author)