What's the point?
It's not 3-d, it's a gimmick. The only reasons for it are to (a) sell more sets and (b) differentiate one broadcaster from another.
As others have pointed out, the brain's depth-processing doesn't match what's being broadcast, it gets broken when the image is in the wrong place (e.g. too close to and too off axis from the eyes), or when other cues break the instinctive action to check something moving fast and apparently close to the head. There's no 'look behind' capability on any of the current proposals - and even if material is shot from two cameras 'just in case' it will never have that. Possibly most significant is the inability to track focus with distance; you're always focussed on the display device irrespective of the image focus point. And of course the small displays have a very restricted sweet spot.
On top of that the whole grammar of film/TV production needs to be changed; you can no longer (for example) play with the focus to indicate an area of interest; fast zooms, tracks, and pans no longer work (you lose the 3-d effect if you do it too quickly), and of course the issue of the viewer throwing up needs addressing.
The data bandwidth required either doubles, or if it remains the same, allows with most displays only half the resolution - which indicates HD bandwidth for not much more than SD display (though this is debatable depending on how you count the bandwidth - but certainly there's less per eye available).
You need to edit everything twice, and possibly shoot it twice, too - you can't just drop down to a flat image if you don't have a 3-d display device. (Well, you can, but for reasons of the film grammar, it looks odd.) And of course the transmission method and the display surface have to be independent otherwise you can't display some things on incompatible displays.
Apart from that it's wonderful. But I can't help feeling that Avatar would have looked as good flat as in 3-d, and without the silly glasses, too.