That, or the Sinclair ZX Spectrum contended memory pages 4-7 which can only be accessed at full speed by the CPU when the display is drawing borders or doing horizontal/vertical retrace.

Toyota already uses QNX - at least in their Touch&Go nav/audio head units sold on european market cars. The OS identifies itself as "QNX Neutrino" and appears to run on OMAP3 ARMv7 hardware (TI DM3730 SoC).

At least some small aircraft have a rotating vane sensor which measures the direction of the airflow in relation to the aircraft, ie. the angle of attack. It may be used to trigger a stall warning if the angle exceeds safe operating limits. I don't know, however, how many larger airliners come with similar equipment - I've seen a picture of one on an A380, at least.

The trick for more sprites on C64 was to generate rasterline interrupts which change the sprite location while the electron gun of the display was still drawing it. This way you could re-use a sprite further down the screen in an area which had not yet been drawn. However, the new location had to be at least 21 (IIRC) rasterlines below the rastercount at which the interrupt occurred for the new sprite to be displayed. Also, you couldn't change the sprite data pointer until the old sprite was fully drawn because weird things would happen.

Oceanic OC-118N?

You know, you really shouldn't believe Jmagic if he says an object has 7800 phong shaded polys...

As far as I know, all the software rendering demos and intros in the 90s claiming phong shading faked it either by envmapping with a prerendered texture of a specular highlight or doing gouraud with a non-linear palette. Both looked reasonably convincing, albeit with somewhat rough and/or distorted highlights.

At least on a ZX Spectrum, the psycho-acoustic models do damage the data enough to make it impossible to load. I exported a few .TZX files into .WAV, compressed them into MP3 and tried to load into my 48k Spectrum from a portable MP3 player. I didn't manage to load the program one single time as every attempt ended with an "R Tape load error". I also tried recording the .WAV onto a minidisc (old MZ-R90 portable) but still got similar results, so apparently ATRAC loses too much data as well. Burning the .WAV files onto CD as audio tracks worked flawlessly, though.