The author provides an interesting survey about the calculus rules for the global approximate minimum \(\varepsilon\)-argmin\( f\) and the approximate subdifferential \(\partial_ \varepsilon f(x)\) (in the sense of convex analysis) of a not necessarily convex function \(f: X\to \mathbb{R}\). Obviously, for the approximate minimum no structure on the set \(X\) is required. However, if \(X\) is a locally convex topological space then close duality relations to the approximate subdifferential can be derived especially by means of the Fenchel conjugation. So many calculus rules (e.g., for the sum, difference, composition, infimum, supremum, marginal function) for the approximate minimum are formulated and applied to approximate subdifferential calculus. A large part of the paper is devoted to the discussion of the inf-convolution and its inverse operation -- the so-called deconvolution. Here, the author extends former results of Hiriart-Urruty, Mazure, and himself.