The Rev controller subtlety article seemed to go into the right direction but ended up missing this completely:
The Revolution's rod controller allows for unified interaction in games! Basically, in a game that goes beyond "kill everything in sight" your character can interact with the environment in multiple ways (like pushing buttons, inserting keys into keyholes, etc, you know, everything Link does). However, these ways are limited and if you don't want a big box of confusing button combinations you will give the player some basic actions that are always applicable and a context-sensitive button.
Now here we got our problem for puzzles: Your player can only do the basic things you've allowed him to do, like swing a sword. If that's not sufficient you use the context sensitive button (CSB). But because the player can't guess that the button's function has changed you'll need to tell him. E.g. the player stands in front of a lever, an icon pops up to tell the player that the CSB now means "pull this lever". With this you can't hide things very well, when the player is in front of the statue with the hidden button you have to tell him that he can use the statue somehow so you've given away that it is not just decoration. Of course you can try adding some stupid text to each statue to mask this by giving a "use" popup with every statue you encounter but that's still not optimal. You simply don't know whether the player really has spotted the hidden button or whether he's just randomly hammering the CSB.
Also, a CSB can have only one function at a time and no matter whether the player really knows what he should do with the element he interacts with you make him do the optimal thing. Say, there was a rock near a cliff. The player wants to tie a rope to it to create a weight for some contraption. He goes there and pushes the CSB. Suddently his character kicks the rock down into the chasm. While that probably was the correct action the player didn't think of it and you gave him a solution.
Or when you have to interact with something that's more complex than just pushing a button, e.g. guiding a crane. In Mario 64 there was a platform with four buttons on it that you could control. But buttons on the floor don't work for everything and most games resort to putting you in "control mechanism" mode and make your controls do different things than they usually do (e.g. guide the crane instead of moving your character). That means a discontinuity in controls, additional complication and usually a limit to what you are willing to implement (because those things would need a special mode in the gamecode).
Compare this to Doom 3: Instead of standing in front of a panel and pushing use, you lower your weapon when approaching a panel and use the mouse to control a cursor on the panel, you can click on different buttons there. It's not perfect but it doesn't put you in crane control mode when you need to control a crane, for example. All panels work in the same way and the player always knows how the panel works.
But with the Revmote you can take this a step further. If the player can always control his hand/equipment with such high precision and freedom you can have him remain in a single state for the entire game. Instead of having a CSB you let the player do what he wants directly. Since the Rev is a nextgen platform it should be able to do the physics for allowing the player to get creative and even find alternative solutions to puzzles. But I'm getting ahead of myself. Since your player now can do things like touch a specific region of any object nearby (please spare me the obvious female character jokes) he doesn't need a message to pop up when he's near an interactive element, he has to spot the element and use it himself. If the statue's left eye can be pushed the player has to push the statue's left eye instead of pressing the "push statue's left eye" button. That allows for much more freedom in designing puzzles. You could have the player use his sword as a replacement for a broken lever, you could have the player reflect sunlight with the sword (as in SOTC except without a separate button so it can be required for one level only), you could let him take a hammer and widen a gap on a wall to reveal a hiden item, etc.
With the Rev the player's character and controls can behave in the same way throughout the game without limiting the actions the player can or has to take in the game. Previous Zelda games have tried to do this as much as possible but they were still limited by the game's action set. With the rod you'll see many more puzzles in Zelda that require new approaches not seen before.
I apologize if this has been written before but there's so much useless speculation in the Rev section that I can't be bothered to tead every single post in this forum

.