A number of people seem to believe that if a certain body of knowledge keeps on changing, then it must be worthless. But at least for science, which tries to understand how our reality works, it would be worthless if it didn't change to match reality. If science didn't throw away models that did not conform to reality, which is checked by performing experiments and seeing if the results actually agree with the models instead of by pure philosophizing, then it's not doing its job as a body of knowledge for understanding reality.
We see this happening in the history of science multiple times, where scientists had to adjust their theories according to what reality showed them, not the other way around. The history of science is made up of a series of approximations that get better with time. While most of the changes in science are small refinements to theories, every once in a while a drastic overturning of existing models is necessary to get a better approximation to reality.
The change of our understanding of light as particles to waves is a good example. In the 18th century, most scientists believed that light was a stream of particles. There was also a competing theory that light was a wave, but this was discounted in favor of the particle theory because while both conformed to experiment, the particle theory had the weight of Newton's approval behind it. If science was like a religion, where some sort of authority is the arbiter of truth, we'd probably still subscribe to this theory.
However, this all changed in the 19th century as new experiments showed results that were more consistent with the wave theory of light. The results of an experiment which attempted to measure the speed of light through refraction and stellar aberration, which Skulls in the Stars has a fascinating account of, were not consistent with the particle theory. Instead, the results were consistent with the wave theory as long as one assumed that the "aether", the postulated medium through which light was supposed to form waves, was partially dragged with matter.
There was also Poisson's spot, which is a dramatic example of how wave theory was more consistent with reality than the particle theory. According to the particle theory of light, a sphere would block the light particles so that it would cast a completely dark, circular shadow. But when Fresnel advocated the wave theory of light, Poisson ridiculed him by showing that such a shadow would have a bright little spot right in the middle according to the wave theory of light. The reason for this is that because the edge of the sphere would all be the same distance from the middle of the shadow, the waves propagating from the edge would all bunch up into a bright spot.
At first glance, there being a bright spot right in the middle of the shadow of a sphere sounds like a ridiculous notion, and Poisson put it forward as such. So when Arago did the actual experiment and showed that there actually was such a spot in a sphere's shadow, that was a really strong indication that the wave theory of light was better than the particle theory. It's a bit ironic that the bright spot is more often known as "Poisson's spot" after the person who thought it didn't exist, although occasionally it's also called "Arago's spot" after the person who actually observed it.
Seeing that reality was better approximated by the wave theory of light rather than the previously widely accepted particle theory of life, scientists didn't just stubbornly cling to the particle theory taught by their forebears. They yielded to reality and accepted the wave theory of light, which eventually merged with the science of electricity and magnetism to give Maxwell's theory of electromagnetism.
And the story doesn't end there. Further developments in the 20th century eventually got us to beg the particle aspect back into our theories of light, and we got quantum mechanics where everything is both a particle and a wave. We also got the theory of special relativity. These two theories were much better approximations of reality than the ones that came before. And it's not unlikely that we'll replace our theories yet again in the future with even better approximations. Perhaps even in the near future with results from instruments such as the Large Hadron Collider, GLAST, or LIGO.