Electrically though, it's the same thing. The output of that MOSFET feeds directly into the 5V rail, and therefore the output of the 5V regulator.
I grant you that from the on-board regulator's perspective there is little difference if the board is being USB powered or powered via an external regulated +5vdc applied to the shield 5V pin, in both cases an 'external voltage' is being applied to the on board regulator, other then the USB path has fuse proteciton.
The root cause of failures some were seeing early on with the Unos (and I bet that the arduino co had some returns to verify bad regulators were happening somehow) may not have been nailed down conclusively. As I said I believe the arduino folks 'official policy advice' to not connect external +5vdc was just being cautious and perhaps to try and cut down on warranty returns.
Perhaps it was just an early bad batch of voltage regulators. I do know that it is a bad engineering practice to 'hardwire' two voltage sources together as in the case if one was powering the board via it's on-board regulator or an external regulated +5vdc applied to the 5V pin, which is the case if the board was then plugged into a PC via USB, hence the reason they installed the MOSFET controlled switch to isolate USB voltage in that case.
Again I am not a fan of the design of the existing auto-voltage selection function if for no other reason is that it takes up a lot of board space and doesn't give me complete control. as in the case if I wanted the board powered from USB even if I have a valid Vin voltage applied. Jumper clips are what I would use.
But the real difficulty is what advice do you give a beginners if he asks about the arduino policy advice to not wire external +5vdc to the 5V pin? just tell them it's wrong advice and to ignore it, or try and explain the historical reason the product page states what it states. I tend to normally not answer such questions unless I'm in the mood to go through all this once again, with no real consensus reached. :~