We study parametric identification of uncertain systems in a deterministic setting. We assume that the problem data and the linearly parameterized system model are given, In the presence of a priori information and norm-bounded noise, we design optimal worst-case algorithms. In particular, we study the interplay between identification tools and nonstandard techniques used in approximation theory. The obtained estimators, called smoothing algorithms, as well as the identification errors are computed by means of the singular-value decomposition of the system model. Finally, the proposed algorithms are tested on real data referring to the tuning of A/D converters.