The classical analysis of the s-process is commonly used to predict the s percentage contribution to the solar system abundance of a given isotope, and by default of the r-process residual (calculated as 1. - s). We discuss the advantages and the disadvantages of this first-order prediction, by comparing stellar model calculations at various metallicities in AGB stars and in massive stars exploding as SNII, with spectroscopic observations of different stellar populations. Observations of short-lived r-process isotopes in the early solar system help to characterize at least three different r-process components.