In case of bigger files to translate and revise, it would be almost mandatory to be able to indicate that segments had been translated by autopropagation and even filter on it..
Imagine you have a huge Excel file, an extract of a PIM system. A significant part of the segments are repetitions. The translator will use autopropagate to make his live easier (he has to give discounts on repetitions).
The reviser however, doesn't have the possibility to see if a segment has been autopropagated or if it was a new translation. The reviser for that reason have to read every single segment. That makes revising even more expensive than translation!
How logical it would be if a revisor could filter out (instructed by the Project Manager) autopropagated segments.