A good theory

I'm reading Richard Dawkins' The Greatest Show on Earth. It's really good--the evidence of evolution that he provides is fascinating and compelling. On page 149 Dawkins writes

a good theory , a scientific theory,is one that is vulnerable to disproof, yet is not disproved.

He gives the example of Haldane's response to the question of what evidence disprove the theory of evolution: "Fossil rabbits in the Precambrian." But I think Dawkins is not quite right. Suppose fossil rabbits did show up and were reliably dated far before the first mammals? Would we stop believing in evolution? I don't think so. We (well, the main problem would be for biologists but all believers in evolution would have to be pretty disturbed) would have a huge problem. But as someone named Dobzhansky once wrote, "Nothing in biology makes sense except in the light of evolution." Fossil rabbits in the Precambrian wouldn't make us abandon the of evolution because we need the theory to account for so many things that we observe. And there's no plausible alternative. So in light of this, I propose a new definition:

A good scientific theory is the most internally consistent and evidence-based framework for making sense of the complex phenomena we observe in the world.