Since I graduated from college with a major in physics & a minor in math, I do understand what you are saying. However, it is still not possible to adjust "mouse speed" (pointer sensitivity, which Apple calls "Tracking Speed") independently of acceleration unless acceleration is zero. Any non-zero value makes pointer sensitivity dependent on mouse velocity, as your examples show.
One group of users do want zero acceleration; IOW, no change in sensitivity regardless of mouse velocity. Thus, the reference to removing mouse acceleration in the title of this topic. This is in fact how the crude, earliest versions of Windows mouse drivers behaved. The result was a mouse that either had to be moved very far to move the pointer all the way across the screen (low distance sensitivity) or very carefully to move the pointer small distances (high distance sensitivity). The solution (which was implemented from day one on the Mac) was to add the velocity sensitive component to the driver, which essentially all modern drivers include.
Another group wants to be able to adjust velocity sensitivity independently of distance sensitivity. This sounds reasonable, but it isn't as straightforward as it seems. It should be obvious that it would make no sense to combine very low distance sensitivity with very high velocity sensitivity, or visa versa. The first would require impossibly steady rates of mouse movement to position the pointer precisely; the second would require impossibly small mouse movements to do so. Neither would be useable.
So it makes more sense for the sensitivities to interact to some extent: as one goes up, so should the other, & visa versa. This rather elegantly eliminates unusable extreme settings & at the same time makes adjustment easier over the range of useable ones. The simplest, most user friendly way to implement this is with one basic tracking sensitivity control & another one that adjusts the scale factor between distance & velocity sensitivity. In effect, it allows users to adjust low mouse speed sensitivity & high mouse speed sensitivity separately without having to know anything about derivatives or jerk or acceleration curves. Since most don't, this is a reasonable approach.
But take this a step further: it makes even more sense not to make the scale factor linear (resulting in the flat, straight line curves you mention). Users typically are not satisfied with flat curves; they are either too steep or too shallow over some part of the range.
Apple's solution was to condense everything into one control that interactively adjusts everything: the distance & velocity sensitivity and the shape of the curve. At the lowest setting, velocity sensitivity is essentially zero; as the setting is increased, the velocity sensitivity increases, as does the distance one. But the shape of the curve also changes. The result, after a lot of testing involving a wide range of users of varying physical dexterity & computer experience, is a range of adjustment that most Mac users find both more than adequate & easy to set.
Obviously enough, some users do not find this range adequate. There is no point in arguing about that. But some I suspect just want to be able to experiment with radical settings even if they aren't really very useable. A few seem to want the impossible, although they my not realize it: a response that is velocity independent, extremely precise, & still somehow covers large screen distances with short mouse strokes.
All I can suggest for those that are not happy with Apple's implementation is to send the company feedback about it, explaining clearly what you want. Hopefully, the discussion about derivatives, interactivity, multiple curves, et al will help with that.
