Bingo !
This is exactly what i wanted HylianTom & others to find by saying they were on the right tracks when they started discussing of it:
It was nearly one month ago, here:
http://www.neogaf.com/forum/showpost.php?p=36316875&postcount=8576
In addition to optimizations of the methods to render a second view of a 3D scene, we can certainly imagine an eye tracking/shape recognition (thanks to the camera that will be able to do facial recognition) technique, to dynamically dispatch the system power either on the TV or on the padlet, to save up the resources for a title having extensive use of both screens.
You must stare at the padlet for a long moment ? Look at it, the system knows that you're focusing on it, blurring/putting in low details mode the action on the TV (but it continue to display something, for you to still be able to localize your character and be aware of the action when you'll switch back to it), freeing quite a bunch of capabilities for the DRC. The same for the reverse case. Thanks to this approach, you can have the best of both world, nearly all the resources available allocated to the screen you're actually seeing for the best complexity & image quality manageable, and still the possibility to display intricate content on the TV & DRC.
The streaming technology used in the Wii U allows that, the time lapse to send what the TV shows to the DRC is nearly instant, so the dynamic distribution of the power depending on what you're looking can be done.