My new paper proposes an infinite-dimensional l1 minimization framework for function interpolation and analzes its approximation properties. In particular, I show that weighted l1 minimization with polynomials is an optimal method for approximating smooth, one-dimensional functions from scattered data.