I don't think their implementation is going to make that much of a difference. From Magnus Manske:

The first article that comes to mind when I hear this is the Seigenthaler one. Didn't an admin go in and make a few spelling corrections to that one? I haven't seen a study, but I would think that most of the worst mistakes will pass right by the average editor. Sure, a page which consists of nothing but "Anthony is poop-head" will get stopped, but that kind of stuff doesn't ruin reputations, it's mostly just embarrassing to Wikipedia itself.

Good point, actually...

It would certainly be better if they could change that default by article-category, or else have some sort of additional "flag" placed on specific articles so that those pages aren't approved until someone clicks a button to explicitly approve them (as opposed to simply making a non-substantive edit elsewhere within the article).

True, but I think it would easily reduce most vandalism. It would also encourage people to be more careful. In the Seigenthaler case, that admin can say they were just focused on spelling and didn't think about the content, but they wouldn't have much of an excuse if they actually flagged it as sighted.

Still, I wonder if it's ever actually going to be implemented.

And then put the real name of the person who approved the article at the top, saying "This article has been approved by X". Then we might be getting somewhere.

Not going to happen, though.

People tend to view stable versions as a simple thing. On off.

The feature should be viewed as a starting point and it should have been designed scientifically. I'm not aware of any designed experiments to verify its effectiveness and point the way for improvement. This is a common problem at wikipedia - there is no scientifically based research and development, nor is anyone empowered to really do this. Stable versions should be tested in many places with design variations by a group that is empowered to revise and refine the feature.