Beside its intrinsic interest for the insights it can give into color confinement, knowledge of the space-time evolution of hadronization is very important for correctly interpreting jet-quenching data in heavy ion collisions as due to parton-medium or hadron-medium interactions. The extracted medium density, critical for determining the formation of a Quark-Gluon Plasma, can greatly change in the 2 cases. On the experimental side, the cleanest environment to study the space-time evolution of hadronization is semi-inclusive DIS on nuclear targets. On the theoretical side, 2 frameworks are presently competing to explain the observed attenuation of hadron production: quark energy loss (with hadron formation outside the nucleus) and nuclear absorption (with hadrons formed inside the nucleus). I will demonstrate that, contrary to usual expectations, the observed approximate $A^{2/3}$ scaling of hadron yield attenuation in nuclear DIS data cannot conclusively establish the correctness of either energy loss or absorption. I will also review a few other observables which may help to distinguish these 2 mechanisms.