I always found peculiar of computer science in general the fact that once something has got good enough, people in the field feel the urge to throw everything out of the window and rebuild the whole thing, just a tiny bit differently, possibly making the same mistakes that have been done in the past.