I think I haven't made it clear enough what my objection is. In my opinion it is the article that pushed the analogy too far. It works great to show that you can't tell what is moving. But when the analogy tries to claim that you can't move, well, it's pretty much begging the question. Of course a floating point can't move, and it says nothing about physics. That's why I assumed some mechanism for trying to cause motion: you can't see if you failed if you didn't try.

I guess I wasn't clear (no sarcasm). My point is based on the idea objecting that an analogy doesn't stand up to being pushed "too far" isn't a useful one; analogies always break down when pushed too far. If they didn't, they wouldn't be analogies, they'd be explanations. Analogies are dangerous ways of thinking, but again, this is generally targeted at people for who that is all they have available.

The analogy is useful, I agree there. The problem is that the article first uses the analogy for something it fits, then uses the analogy again for something it doesn't fit. The conclusions drawn from the second use are invalid. I don't object to the analogy, I object to that second use.

I had the same objection as the op actually - if you wake up at "rest" with just yourself and a bed, and accelerated away from the bed, you can in fact know whether "you are moving" or "the bed is moving". It is because one of the two have only been in 1 reference frame, while the other has been in 2.

Basically the same issue as the twin "paradox". I can appreciate the assumption of a constant velocity on wake though, as a way to simplify the story.