"The most immediate problem is that this data is stored in an easily-readable form on your machine. Any other program you run or user with access to your machine can look through it."

Apple may not upload it while syncing or by using a scheduled cron job, but any single individual app can read it. Also, as the others said, prove to me at no event does any proprietary apple application access the file. The location data resolution is set to one second intervals, that is insane. They can easily know when I take a piss just by how often I frequent that specific location for short time periods.

The 2D to 3D conversion process has really destroyed the reputation of stereoscopy. It is a poor attempt at achieving the real thing. It can only be done with two separate cameras side by side filming the same scene (or two digitally rendered scenes with separate eye points). We have two eyes, so cameras require two lenses to perceive the scene that is being captured/recorded. The replication/presentation on a video screen is another challenge.

I'm not sure that most people understand that you are suppose to look into the screen, as opposed to having things come out of the screen.

I love perceiving depth in movies, and once head tracking is used for modifying the view frustum dynamically, and it is combined with stereoscopy, the 3D experience it going to be incredible. Unfortunately, this is impossible for more than one person viewing at the moment. Unless they can somehow achieve a massive enough frame rate for multiple viewers (and shutter glasses set at different time intervals). Though by the time this becomes practical, I'd say light fields will have taken over. But even with a light field, we will still want infinite backgrounds (and side views) with depth perception. I'm not sure that I really like the idea of the parallax barrier used in the 3DS, but I imagine it produces satisfactory results.

I imagine that people actually have trouble with convergence. I'd say that it is the same when they try to view a stereogram. I have a lot of trouble viewing a stereogram, but I have no trouble with viewing 3D video. I think that people incorrectly squint their eyes too hard as if they are trying to see a far off or blurry scene, which would give anyone a headache. It is a natural reaction for people to squint their eyes when they are trying to perceive something, especially for anyone who wears glasses.

Binocular cues are incredibly important for a true viewing experience. Tennis has already become huge adopters of the technology. It allows them to tell whether a ball is inside the line. Soon all the other sports will be the next to follow. I can't wait to look back and laugh at all the haters/flamers.

Currently, the Nvidia Tegra 2 chip has 8 cores. A high powered dual core A9-cortex ARM chip, a low powered A7 (for idle state and handling other low power interrupts), a core for HD video encode, a core for HD video decode, a core for audio, and a GPU. Though it lacks the Neon instruction set for full performance SIMD.

OpenCL is currently on its way into the mobile world. Soon the mobile world will also make the transition into streaming multiprocessors. The thought of holding back these innovations is just ludicrous dribble. MIMD is soon going to replace rasterization with backwards rendering, which will require a high amount of complexity, which a quad core would be more ideal. Especially, if you want to exploit the heterogeneity of OpenCL with both the on board GPU and CPU. Maybe cloud rendering will eventually replace this, though I have my doubts. I want to be able to render my screen locally without connecting to the internets.

I hope to one day be able to run test simulations on my tablet while I am waiting in line for lunch, see the results, and then execute heavier distributed processes.