OK, now you are just being deliberately dim. Obviously exactly the same thing could be said about how well the wall is attached to the ground.
Right or wrong do you actually want a rational discussion, or to deliberately be a pain in the butt?
I am right, and I
and trying to have a rational discussion, RossFW!
When the cart is taken from windspeed, it is mentally "box-cared" to show the relationship that you say exists.
It is agreed that the boxcar is a reduced environment to show the "important" effects? Yes, but when on the belt, you need to restore the connections that
sustain those effects.
We can say the velocity requirement is met by the belt's motion, so that leaves the KE. You agree the cart's KE w.r.t the ground is zero. To be a valid model, it must be there, "relative" to something. The only candidate is the belt.
How is it possible that the belt can do this? Simply the speed is not enough, because the mass must be included to make the relative relation correct.
I can see know means of calculation, other than to simply say it "must" be so, yet it cannot be measured to sustain that claim.
If you take a KE measurement from the belt, it will not be correct. How can you do that, without causing the cart to loose it position of zero KE w.r.t the ground? Like the runner!
I have tried to do this test RossFW in anticipation of your objections, yet it always fails. Even moving the cart a little from an imaginary vehicle will fail, because the cart will gain that KE relative to the ground
and the belt.
(That is why the fork moves the cart)
This is a least a part of what I have been trying to tell you. Zero KE objects cannot interact with the environment, yet remain at zero KE relative to that environment. True, but of no use.