I'm currently re-reading Philip Kitcher's excellent book Abusing Science: The Case Against Creationism. Although this book came out some years before the evolution of that new form of Creationism called Intelligent Design, this book is nonetheless still very relevant to the arguments heard in the creation/evolution debate today.
But, there's one thing that always rubs me the wrong way when I read descriptions of science by philosophers (Kitcher is a philosopher of science; in fact he's my favorite philosopher of science - I highly recommend Science, Truth, and Democracy, and In Mendel's Mirror). Philosophers are often so concerned about debunking a naive scientific triumphalism that can be found in some scientists' writings, that I think they distort the picture of scientific progress somewhat.
Here are some quotes from Kitcher (p. 34) "But scientists often forget the fallibility of their enterprise. This is not just absentmindedness or wishful thinking. During the heyday of a scientific theory, so much evidence may support the theory, so many observational clues may seem to attest to its truth, that the idea that it could be overthrown appears ludicrous... Trained biochemists will talk quite naturally of seeing large molecules, and it is easy to overlook the fact that they are presupposing a massive body of theory in describing what they "see"... No theory in the history of science enjoyed a more spectacular career than Newton's mechanics. Yet Newton's ideas had to give way to Einstein's... [Scientists'] enthusiastic assertions that evolution is a proven fact can be charitably understood as the claim that the (admittedly inconclusive) evidence we have for evolutionary theory is as good as we ever obtain for any theory in any field of science."
As a caveat, I may be reading more into this than Kitcher intended, but I'm going to get on my soapbox anyway. The impression that readers might draw from this passage is that theories like relativity, evolution, quantum mechanics, plate tectonics, thermodynamics, etc. can go the way of phlogiston and other discarded theories that you only hear about in a history of science class. However ludicrous it may seem today, a future generation may look back and see how wrong we were (which is exactly what intelligent design advocates believe). Kitcher's response seems to be (in a later part of the chapter, not the passage I quoted) that yes, these theories may go the way of the humour theory of human physiology, but they are the most well-justified theories we have by the facts available to us today.
I don't think that does our current ideas justice. Yes, Newtonian physics had to make way for relativistic physics, but essentially every college science major begins physics with a whole semester of nothing but Newtonian mechanics, and then another one devoted essentially to Maxwell's theories of electromagnetism.
On the other hand, you don't start freshman chemistry with a whole semester on phlogiston.
There's a difference between these two examples that philosophers frequently downplay or leave out in their descriptions, and I think it gives people the wrong impression.
I'll make a prediction: just as we still teach a lot of Newtonian physics, 100 years after Einstein's miraculous year and nearly 80 years after the first development of quantum mechanics, we will also, 100 years from now, still teach basic evolution to every college biology major, we will still teach the basic ideas of genetics and molecular biology, and we will still teach the basic ideas of plate tectonics. These theories may be superceded by radically different ideas; still, like Newtonian mechanics, the basics of evolution, molecular biology, and plate tectonics will still be there and relevant. No one should be under any illusion that evolutionary theory will look like phlogiston in 200 years. Science does make progress, and barring that the laws of physics miraculously change at 12:05 am on January 12, 2087, our most solid ideas in current science are destined to stick around, in some form.