Extracting nucleation rates from current–time transients: Part I: the choice of growth models