Seven historical cases are given in which physicists agreed, in the long run, that major progress had been accomplished through a new theory or within a given theory. These cases involved conflicts between formal and empirical criteria of progress, with the latter kind, empirical adequacy, winning in the end. The physicists’ ability to judge and compare the empirical adequacy of their theories crucially depended on a modular structure that defined a sound basis of comparison. A definition of physical theories is proposed in which this structure plays an essential role. The concomitant notion of progress as an increase in the modular measure of empirical adequacy is compared to a few philosophers’ approaches to scientific progress.