Now that we've all made the leap to watching and listening to content delivered via a computer (whether we realize it or not) I'd be in favor of ANY format that's open ended, so that it can be extended without huge costs in ancillary equipment.

It's my preference that the content should eventually be delivered in as much resolution and detail as possible and the 'best fitting' to the final output device should happen at the end. This is why I really like the concept of getting free of x.y audio too - just encode the mixing engineers intentions into the data stream and let the local decoder sort it out at my end.

This will probably happen, someday, but I'm sure not holding my breath on that one.