An algorithm is given for determining an optimal $b$-step approximation of weighted data, where the error is measured with respect to the $L_\infty$ norm. For data presorted by the independent variable the algorithm takes $\Theta(n + \log n \cdot b(1+\log n/b))$ time and $\Theta(n)$ space. This is $\Theta(n \log n)$ in the worst case and $\Theta(n)$ when $b = O(n/\log n \log\log n)$. A minor change determines an optimal reduced isotonic regression in the same time and space bounds, and the algorithm also solves the $k$-center problem for 1-dimensional weighted data.