We have to scale this by the inverse of the t distance?

[sirop]

Came across

//tests and such have shown that we have to scale this by the inverse of the t distance, and i'm not sure why
//intuitively it would just be the 2nd derivative of the position function and nothing else
//if you know why please let me know

My uneducated guess:
https://stackoverflow.com/a/23980479/4599792 , 3rd paragraph:

One caveat is that the above tangents are computed for the interval (t1,t2), so if you want to evaluate the curve in the standard interval (0,1), simply rescale the tangents by multiplying them with the factor (t2-t1).

Do we have here exactly the opposite case: tangents are computed for the interval (0,1) and therefore the necessity to "to scale this by the inverse of the t distance"?