Sunrise and sunset points are where the earth obstructs the sun, are they not?
That's what Pete was saying too but I don't agree... in EA, sunset is not occuring when light is obstructed by the earth, sunset is occuring when light is curved upwards away from the observer. There are many rays that do travel downwards and hit the earth, just like any observer at any time of the day will see those light rays bending down to them. We know the earth obstructs that light, because things on the earth are illuminated. But at sunset, those rays of light, according to EA, are curving so much as to be tangential to the surface, so the sun appears in the distance "level" with the horizon, and then after that time, because the light is now curving upwards away from the observer, then the sun is no longer visible. So it's not invisible because the light rays are being obstructed, it's invisible because the light rays are not reaching the observer.
Any observer of this light will consequently see the light approaching from below.
No, this is quite simply not the case.
How is this not the case? Can you please draw a diagram?
Why is it impossible to see that light? Look at the sketch above, if you go to a place where the upward bound light is going, why is it not visible?
Because the light rays that are actually relevant are both more numerous and luminous.
So the relevant light rays, as you mentioned else where in the post, are the ones that actually strike your image plane e.g. your retina right? So a rough explanation of how a lens works: it "ignores" all other incedental light rays, and only the light rays hitting the lens from a limited number of angles will end up passing through and being focused on your retina.
So because we know that your eye (or camera) is selective of the light rays that it actually "sees", then we can be selective in the light rays we draw right? The light rays that people are drawing, travelling downwards and then upwards towards an observer, should be entirely relevant: we know that a light ray we draw in the direction the observer is looking, is going to be seen. And light rays at an angle can also be seen: we know our field of view is around 60 degrees yeah? Any light rays entering from within that field of view, and travelling in certain directions that the lens will "bend" and focus to a single point, will be seen.
Having said that I take Tom's point, and have mentioned that previously: the scales involved are huge, so our not-to-scale drawings can not really "prove" how much of the sun should be visible below the earth and at what distances/angles etc. I'm thinking without a mathematical formula for the curving then this would be impossible, but perhaps we have sufficient information to provide a "rough" idea within certain constraints?
For me though, I think it's more important that we all agree at least on what seems to be the central tenant of EA: "sunset occurs when light no longer reaches the observer, because it has been pulled upwards away from them"
Regardless of whether or not that effect would actually be seen, then it will to my mind also require the sun to "disapear" from the top down. The top is higher than the bottom (!), so the light rays from the top will curve upwards away from the viewer before the bottom light rays. Trying to bring an argument of "oh well there's heaps of different angles of light rays" is irrelevant: we KNOW that only certain rays will be focused by our eye, and we know that focus mechanism is the same for the top of the sun as the bottom of the sun, so all things being equal, the top would disappear first, the bottom would never "reach" the horizon (aside from any light curving back upwards to the observer). The sun would never appear "half set" over the horizon: if the bottom rays arn't visible, the top rays would not be visible either.