1. The Paracorr's correction of field curvature is slight, and is related to having the focal length lengthen 15%. this makes field curvature in the telescope less, and since field curvature in the eyepiece and telescope are additive, there is some correction of field curvature.
2.Bratislav's idea that a short f/ratio mirror contains rays of all f/ratios is dependent on the idea that some rays strike and use only a part of the mirror. Unfortunately, and I'll explain what I say in a minute, EVERY RAY HITTING THE MIRROR USES THE ENTIRE MIRROR BECAUSE EVERY RAY STRIKING THE MIRROR IS A WAVE THAT HITS THE WHOLE EARTH. That means that, wherever the star is in the field of view, the f/ratio is the same. What does happen is that the front of the tube and the secondary size do vignette that wave, reducing its brightness from an axial wave, but the wave itself is focused on a point at exactly the same distance from the mirror. If you'd like an illustration, stand 1' from a wall with a bright flashlight. Shine the flashlight on the wall straight ahead. The light on the wall is circular. Without moving the flashlight toward or away from the wall, angle the flashlight 45 degrees to the side. What shape is the light beam hitting the wall? Oval. Now, this is an oversimplification of what occurs on the focal plane of a paraboloidal mirror, but it is somewhat illustrative of why coma is worse farther off axis.
3. When the wave passes through the Paracorr, the outside of the wave is bent by the edge of the lens. Like any light ray passing through a lens, the angle of incidence determines the angle it exits the lens. All f/4 telescopes, regardless of focal length, have the same angle of incidence entering the Paracorr lens. Accordingly, the lens curvature can be designed to fully correct a particular angle of incidence (f/ratio). But what happens when the f/ratio is longer? The rays that hit the lens do so at angles closer to parallel, which makes the angle of exit warped to a greater degree. So, if the goal is to reduce the size of the comatic star image to no larger than the size of the Airy disc, this will happen over a small range of f/ratios. At very long f/ratios, where the rays hit the lens at nearly parallel angles, the warpage of the lateral rays can bend the ray the other way and cause "reverse coma". At what f/ratio this would occur, I don't know, but I wouldn't try a Paracorr in a Schiefspiegler.
4. There is a range of f/ratios that would be completely corrected. On the Paracorr, that would appear to be f/5.5 to f/8. That says that Al chose to make the curves gentle to avoid spherical aberration. The 15% increase in focal ratio was necessary to avoid any detectable spherical aberration from the lens for on-axis images. This is also why the other commercially available coma correctors DO exhibit increased spherical aberration, though this is not as important photographically due to the larger size of the star images compared to visual use. It also means that if TeleVue had chosen to fully correct an f/4 telescope, spherical aberration would have had to be controlled some other way, such as additional lenses or aspherical curves, both of which would have made the Paracorr an unaffordable luxury for a chosen few, as well as much heavier.
5.The point to get across, here, is that the linear size of a comatic image will be exactly the same size 15mm off axis in any scope of the same f/ratio, whether having a 25" focal length or a 100" focal length. There is a difference of image scale, though, and it depends on whether you look at coma from an angular or linear distance away from center as to how you describe it. I suspect Bratislav and I were talking about coma from opposite points of view.
But when we look through an eyepiece, we are looking through an eyepiece with a set linear field stop, and the coma present in the image at the edge of the field of any eyepiece is dependent on the linear size of the field, and it is dependent on the f/ratio of the scope.