No science is not "merely a structure for generating beliefs". Rather it is an approach for distinguishing among "generated beliefs" (ie hypotheses) and discarding those that do not fully explain or are unable to explain quantifiable observations.
Ouch, my hairs, you have split them! I agree, but the point still remains that science deals with belief, not knowledge or facts. I reiterate: it's easy to delve quite deep into epistemological arguments to claim otherwise.
The fact that many great scientists have been religious men with "sober approaches to faith" is hardly material to the difference between science and religion. Science is all about HOW WE KNOW, not what we think we know. Good sophism is never a substitute for science. Neither is religion.
Oh boy, I think you've got me all wrong. Assuming a modicum of philosophy, I'm trying to say that there's less of a difference between science and religion than we'd like to think. They are both systems by which we establish belief and they are both prone to error and misinterpretation. Descartes and Newton, for instance, both created extensive philosophical and religious writings which mutually benefited their theories of mathematics and physics. Scientific theories are a collection of discrete data points, but we live in a continuous universe. Dare I say, to be a good scientist, it takes a certain amount of faith-based value to fill the gaps between and beyond the data, of which there are an infinitude. For the atheist, that value might come in the form of believing science will limit to absolute knowledge. For the theologian, that value might come in the form of believing in the existence of an omniscient being.
As far as your "fun philosophical question" is concerned, neurobiology tells us that there are all kinds of stimuli our nervous systems perceive and interpret inappropriately all the time and that there are often many ways and circumstances in which the "brain" can "fool itself". However, none of this is really material to the question of which is more valuable to humanity, science or religion. Excellently crafted sophism is no substitute for science.
...nor is good sophism a substitute for thinking for yourself. I never mentioned or alluded to "which is more valuable," so, once again, you've completely misread. Whether our nervous system is prone to "fooling" us or not, there's only so many things, 4 to be exact, our physical bodies can directly interpret: visible light (sight), audible sound (hearing), appreciable physical interaction (touch, including heat) and significant chemical concentrations (smell/taste). Everything else stringently requires we engineer methods of translating into those 4 media. My question posits that, if there is something which we cannot actually perceive, neither directly nor indirectly, science would not be able to extract beliefs about it. Dark energy, for instance, although not exactly entirely imperceptible, has no real perceivable quality; it's just a massive fudge-factor for a discrepancy in our predictions of the universe's expansion rate and the data. Eventually, our theories and paradigms will shift enough to accommodate this discrepancy. But what if there's something that truly is imperceptible by our senses? If we can't perceive it, we can't take data. If we can't take data, we can't use science to refine our beliefs regarding it. That is the question I'm proposing.