Price is most important. If Nintendo goes in higher next gen, they'll cut off a sizable percentage of those who picked one up because it was the thing everyone wanted at the price most were willing to pay.
1:1 control would be the best feature they could add (I see HD display, media playback, online functionality all as a given). The engineering challenge comes in making it just as simple to use without making implementation excessively difficult for developers.
Imagine something like a motion capture suit (impractical for a mass consumer device due to sizing, cost, etc.) or even just a pair of gloves, booties (again impractical due to sizing) or maybe just a set of adjustable belts (wrist straps, waist or chest strap, head band, ankle straps) with motion capture sensors on them. Sounds good so far. Easy and relatively cheap to manufacture. Console comes with a front mounted camera to capture motion sensor movement. Still good cost wise.
Now from the developer's standpoint. How do we implement all that sensory information allowing for fully independent movement matching on screen animation and action? And do it cheaply?
Well... actually it can technically be done. It's just the cost/difficultly issue.
Using proxy motions that are close to character animations that the player pantomimes could easily be used. But then it's an issue of how much action can the developer make the player execute to play through a typical game session, and how much of a market is there for people who really want something this sophisticated when most are happy with two hand held controllers that create on screen action when moved.







