But there's the assumption, isn't it? Actually two: that reusable will be cheaper, and that reusable and reliable can go together.

I'm certainly not saying they can't -- I am, quite literally, not a rocket scientist -- but the Space Shuttle program was the triumph of the idea that "Of course it'll be cheaper if we can re-use the vehicle." Which didn't work out in practice for the reasons outlined upthread. My point is that a space program should pick the goals of "This should get people up and back safely and cost $X a launch" because those are measurable. Maybe the best way to do that is with a reusable vehicle! But to make reusability a goal in and of itself, due to a sentimental attachment to the kind of vehicles we used to travel by sea, seems silly.