# Remez algorithm implies that the exact value of the function is known?

I'm reading up on the Remez algorithm(s), and my understanding is we are minimizing the maximum difference between the true and approximate function (over a certain interval) by iteratively changing the coefficients of the polynomial that approximates the function.

However, by doing so, don't we need the exact value of the function, for example at the control points? If so, then would that not defeat the purpose of the algorithm, since often times we approximate the function because we can't evaluate it directly?

Please let me know if my understanding is correct. Thank you.