This is an alright article, but I think it misses the conclusion that arises from the whole of the study. Their argument is that perception of return trips as shorter has to do with expectation of duration not familiarity with landmarks...they do say that trips we are familiar with (like commute to/from work) don't have the effect because time expectations are appropriate.
The problem comes when perceiving travel to/from less desirable locations, or in a case like a distance run with fatigue on the second half, where they found the situation to be reversed. It seems clear that expectations should be the same for both these instances and the "standard" one but the time perception is the opposite. This should imply to them that perceptions of time are related less to the journey and more to the destination, i.e. some anticipation of arriving. Positive anticipation would stretch out the journey and negative shrink it down. Neutral anticipation should only show an effect if combined with a positive or negative from the other leg.
Now, figuring out positive/negative/neutral anticipation is a harder thing, but it seems that it is less an expectation of time and more an anticipation of being at the destination that causes us to distort our perception of time.