http://www.viatech.com/en/boards/
I run a couple of servers using VIA processors, they consume very little electricity, I run storage, web and mail servers on them. When Intel came out with the Atom, it used a little bit more electricity but was significantly more performant than the VIA. VIA has had new processors since but I don't know how they compare to the Intel low-power line.
Extremely ugly first day prototype of that plasma cutter controller, the VIA board sits in the hard drive enclosure on the right hand side, you can just about make it out.
What did you build?
https://www.youtube.com/watch?v=dAyBz8XIMsU
http://www.silverball.net/images_martin/tech_pc_case.jpg
It was a MediaGX 5520 (233Mhz) motherboard with a custom PCI board holding an array of 8mbit ROMs + flash memory and a DSP for sound playback. Douglas Comer's XINU was used as an operating system and ran from ROM.
The "too early for it's time" was less about the pinball machine and more about the MediaGX. A small Geode board + Linux could have been a really nice platform. I think a lot of ATMs are actually using this formula.
I've worked on some coin op stuff as well but more video game oriented. The original software ran on Windows, I did a linux port to show that 'it could be done' without windows. The hardest part was to switch from one game to another (multi game console) without a video mode reset.
Another application for that board is on-board computer for cars. I've got a very similar board in a nice little case that just fits under the passenger seat and wires into the car by running cables to the side of the transmission tunnel. Boot times are short (runs off a flash disk) and media are stored on a small 2.5" hard drive. I like these little systems.
Anyway, I'm surprised they opted for a single threaded, non preemptive system.
Well, they should be, but apparently they are not:
http://www.telegraph.co.uk/news/worldnews/northamerica/usa/1...
Whatever you could conclude from that article the bug bounty clause ruling those systems out would suggest they are indeed connected, as does the FBI statement (which appears to have taken a part of the conversation out of context but is in fact later on verified by the wired people as having some basis in fact, other than that the guy did not actually control the aircraft he should not have been able to get as far as he did).
I agree with you those connections really should not exist but there seem to be at least some wires bridging the air-gap, maybe read only, maybe not depending on the kind of plane.
Fortunately there appear to be enough safeguards in place to limit the damage that could be done and from what I understand these have held up. It also seems that the 'hacker' is a bit of a bragger but he got in a lot further than they initially gave him credit for, and in a way that is so simple that it makes you wonder what else they missed. The safeguards now seem to hinge on knowing what commands to issue and how to bypass authentication but that is stuff we see in an internet context on a daily basis so that seems to be - to me at least - very thin ice to skate on.
An airgap would be better and simpler.
Enfin, they presumably know what they're doing, personally I'd prefer for such a bridge to simply not exist.
If he had not made these stupid claims it would not have made me feel any better about all this.