I think you answered it with point 1- theoretically you could swim along the AP line wth GF of 0 the % only is relevant once you go above that so the GFlo is a moving target so to speak and not a vanishing pointAwesome! Here's one more thought that just occurred to me, since we are having such fun with this graph.
Unlike the M-value line and the ambient pressure line, the ceiling line doesn't really represent any physical phenomenon that you could measure. It's just shorthand for saying "I'm making my first stop at 30%, I'm getting out of the water at 80%, and in between those two dive waypoints I'll just interpolate in a linear fashion to determine intermediary stops". But in this diagram, we have the line extending to the right of the first stop. Which raises some questions.
1) Does the line really exist before (deeper than) the first stop? I don't think that it does, because it is defined as having maximum depth at one end, and the surface at the other. If you make your dive deeper, then the algorithm is going to generate a different line, as opposed to tracing backwards along the same line.
2) What do you call the point where the ceiling line intersects the ambient line? Mathematically, if you extend them both to the right, they will cross somewhere, since they have different slopes. But given the fact that we represent the slope of the decompression line (blue) to be shallower than that of the ceiling line (green) during the ascent between stops, this intersect isn't at the point of ascent from maximum depth (at least, the way that it is drawn). I think that this intersection doesn't actually exist, because the ceiling line has no meaning beyond the depth range of the dive that generated it.
3) Complicating question #2 is the fact that the slopes of the M-value and ceiling lines are arbitrary, since there are no units on the graph, and the diagram is drawn to illustrate a concept rather than provide an accurate measurement. Even if you were to put PPN2 numbers on the X and Y axes, this diagram is just a simplification of 16 different compartments, each with different M-value lines (and even the 16 compartments are an arbitrary construct).
I'm not a mathematician, and I may be missing something here. What do you guys think?