Realism is as much an assumption as non-realism. It doesn't feel that way, because realism fits the common sense of some evolved monkeys. Common sense sometimes is correct, sometimes is not. It doesn't really mean anything when we deal with these matters.
In fact, I would argue that Occam's Razor advises us to prefer non-realism. With non-realism, we are restricted to something we know to exist, by virtue of our human experience: our first-person perception of things. The independent reality, or third-person perspective, is undoubtedly a useful model, but we have no empirical data to suggest it is real. In fact, I would argue that quantum mechanic does raise some serious issues w.r.t. the this model.
Btw, this debate is almost as old as human culture. It was already present in Aristotle (materalism) vs. Plato (idealism). You can also find many traces of it in Eastern thought, e.g. the idea of Maya, or the question of: how can you really know if you are dreaming right now?
> not somehow a proven component of reality
There are no proven components of reality. Proof is something possible in the domain of mathematics, if one accepts certain axioms. Science is not in the business of proving things, it is an empirical endeavor. It gives us theories that resist falsification (until seen) + eternal doubt. Once you abandon doubt, you abandon science.