I agree that publishing negative results would help the whole enterprise of science, but the prisoner's dilemma is pretty clear. I would guess that the authors have been "trained" well enough by publication incentives that they didn't bother to highlight limitations even in their original manuscript. That kind of information shows up in the paper only by absences and implication; it's only informal chats with peers where people talk honestly and thoroughly about the negative results and limitations of their approaches. It's a shame that scientific publication ended up this way. It's one of the reasons I left research to write software.