We systematize the study of dynamic multiscaling of time-dependent structure functions in different models of passive-scalar and fluid turbulence. We show that, by suitably normalizing these structure functions, we can eliminate their dependence on the origin of time at which we start our measurements and that these normalized structure functions yield the same linear bridge relations that relate the dynamic-multiscaling and equal-time exponents for statistically steady turbulence. We show analytically, for both the Kraichnan model of passive-scalar turbulence and its shell model analogue, and numerically, for the Gledzer-Ohkitani-Yamada (GOY) shell model of fluid turbulence and a shell model for passive-scalar turbulence, that these expone...