i sort of see what you're getting at, tom, but you're still not getting the explanation quite right.
you're talking about it as if i'm saying that the sun appears to be in a different location than it really is, but that's not what i'm saying. i'm saying that the straight line path between the sun and the moon does not appear straight to us. the moon and the sun are where they appear to be, but the spacial relationship between them is distorted from our vantage point here on earth.
so, for example, the 90 deg intersection of the wall and the ceiling in my room makes a straight line, and it connects two corners; it forms a straight line path between two points. but this straight line path doesn't appear straight to me. now, the illusion is not that the corners of my room look lower than they really are. if i walk toward either corner, i have no trouble finding it. it is where it appears to be. but my perspective of the world around me is not perfectly euclidean. again, sections 3 and 4 of the paper i posted describe all of this in more detail and probably better writing.
in other words, you're making it sound as if the explanation is that these objects appear in different locations than they truly are in euclidean space. the actual explanation is that the way we see things is not a perfectly euclidean representation of the things we see.