Brainstorm

The main camera of the device (the one with lots of megapixels) should be capable of capturing depth information at least in the same range as it can focus, with 30 FPS minimum, and depth resolution of at least half a milimeter.

The secondary camera should have similar capabilities, but with resolution and range reduced if necessary, but still allowing project Natal like functionality from across the room as well as close range head and eye tracking as well as face recognition and expression recognition.

Bonust points for photo and video recording software that fakes two camera stereo recording by displacing pixels based on the depth map; extra bonus for image processing to compensate for occlusi

The technology already exist, i dunno if it can reach the required performance though.

Similar to the ideas from http://www.youtube.com/watch?v=8ENfPYpkHp4 , with the required software included, and with video recording at a minimum of 30 FPS

Any surface of the device that is visible during normal use (basicly the outside) should be a screen, inlcuding keys and buttons, areas that need to be transparent, such as over cameras and other sensors would have screen technology that besides RGB also has A, transparency, when fully transparent shuold be just about indistinguishible from plain glass.

The whole surface, including the screen and buttons, can detect the distance to the closest solid object, allowing for touchless finger gestures and even small scale depth scanning (perhaps even to the point of reading fingerprints of the hand holding the device)

One way this might be done is by having an array of  tiny pulsing IR LEDs using time of flight to measure the distances (LEDs can be used to measure light besides emitting it)

Like on a desktop PC, it should allow you to install all sorts of Linux, and even other OS's, without the need for special purpose OS

The device would run linux, with the smallest amount of binary blods as the law allows

With multiple microphones it should be possible to filter a sound source by it's position in space, sllowing background noise to be filtered out and your voice to be recorded even in noisy environments.

A while ago i saw on TV these speakers that were an array of tiny ultrasound speakers that could send audible sound in a straight line over a significant distance without dispersion, it works by modulating ultrasound with audible frequency sound, because ultrasound is higher frequency than audible sound it can travel a bigger distance in a straight line without spreading and thinning out. If you bounce the emission of one of those on a wall it would sound like the sound is comming from the wall and not the speaker, if you aim it at one person, the person next to that person would not hear it (of course in this case it depends on the volume)

Laser projectors don't need refocusing when projecting at different distances and allow forhigh resolution and high contrast projection (obviously contrast is dependent on environment lighting)

I propose a 1:1 apsect ratio so any aspect ratio image may be used just by adding black bars on the sides or up and down when necessary

An example of the technology:

http://www.youtube.com/watch?v=7UfarRM0BoM

Each piece has it's pros and cons for each type of usage, so the combo is best for total flexibility

So we can make it better bit by bit instead of having to wait for a whole new device to be launched, and competition will give us better hardware for better prices, and since it's open home made additions are also possible.

Place two or more side to side, and you got double the screen area, with relative orientation recognition, even if they aren't aligned, the image will still fit together,  should be compatible with other models too, witht he image adjsuting for the different screen sizes. Also with the relative orientation recognition, 3D audio should be distributed appropriatedly across the devices, taking in consideration each device's audio capabilities as well (don't try to dedicate a tiny thing with only a tweeter for the bass components for example)

The device should not require special pockets in order to be carried safelly, and should be confortable to hold with a single hand (without comprosing usability with two hands)

There shouldn't be need to purchase additional coomponents to avoid scratches, it should handle faling and draggign against hard and sharp surfaces from factory, and of course shuoldn't ever get noticeably scratched from normal use.

Even for power users free space should never be an issue, but means to use external storage should be provided for practivity among other things.

The frequency range should contain all allowed frequencies in the world, with restrictions applied by the OS for regional versions (so if over here i can use a frequency you can't over there, with the same device i can use it and you can't)

Including frequncies like airtrafic and police radio so people that got authorization to listen to those frequencies can.

This should allow for example for people to use the device together with purpose made walkietalkies with full compatibility, and even pick up TV broadcasts and decode them to watch