And I seriously fail to see the 'for its own sake' here. This cleverness reduces the number of moving parts, does the 'dont repeat yourself' and thus reduces ways modifications of the code could break it due to overseeing a case.
I just shuddered to think how this would be done in java - with a LinkUpdater interface, and a RootLinkUpdater and ElementLinkUpdater, and one instantiation per iteration, and suddenly we have a lot of code around to keep the active code's function simple - to a point where it is completely pointless because of adding a lot of machinery. (And java does collections differently anyway.)
When I look back on my career, often when I've said something is "clever for its own sake" was only clever because of a lack of understanding on my part, and rarely for its own sake, outside of toys and obscurity competition where such cleverness is the point.
One argument is that code should be written so that junior or average (or even below average) developers can be put to work on it. Another view is that part of a junior developer's learning experience ought to include not shielding him from alleged complexity or "clever" constructs, because sometimes they really are better or even unavoidable.
Cleverness for its own sake is one thing to avoid, but I think that's almost too subjective a standard to use.