In the GS/OS Reference Manual, Chapter 10
is about Handling Interrupt Signals.
If none of the interrupt handlers accepts a given
handler it becomes an unclaimed interrupt.
Seems pretty strightforward right?
But there are possible causes for unclaimed interrupts
and I kid you not. This one caught my attention:
random transient phenomena such as cosmic-ray or
subatomic particle bombardment
They're kidding right?
Is this Apple's sense of humour?
Or can anyone with a IIGS verify this claim.
Yup. Very true. Millions are spent designing specialty computers and shielding to guard against such events.
It is why NASA install shielding in their spaceborne systems. And it is why some spacecraft have multiple computers working in parallel, to have each other check results. It is a limiting factor in CCDs in digital cameras in space. And memory systems. Especially flash and DRAM.
I could go on and on about this. But I'll leave it to others to cite more real world examples.
Yes. Electronics are vulnerable to cosmic rays and other high-energy effects. That is why servers use 9-bit parity memory, and why the military requires MIL-STD electronics which are radiation hardened:
https://en.wikipedia.org/wiki/Radiation_hardening
Story: while working at Datasoft around 1987, my coworker Rick Mirsky had recently completed the Apple II port of the arcade game Mr. Do!
While in between projects, he pulled out all the chips out on his personal Apple II (not sure if it was a II+ or //e), and replaced them MIL-STD radiation-hardened chips. He joked that he was ready to program in space, and if the US ever got hit by an electro magnetic pulse (EMP) attack, he'd be the only person with a working Apple II.
-JB
@JBrooksBSI
It is indeed a real thing, but realistically it's not likely to be much a problem for an Apple IIgs.
Most high-energy cosmic rays with the potential to flip a bit are stopped by a combination of the Van Allen radiation belts and the upper atmosphere, but some of them *do* make it to the ground and contribute to "soft errors". A study by IBM back in the 90's suggested that the mean cosmic-ray-induced soft error rate one might expect on a 256 megabyte memory array would be about one flipped bit a month, and that error rate may actually be higher with modern electronics, because the likelihood of a cosmic ray succeeding in "flipping a bit" in TTL logic is proportional to both the intensity of the particle and exact details of the circuitry in question, with factors including the "feature size" of the silicon gates, the switching voltage, cell hysteresis, etc. Basically, the smaller the cell and the lower the voltage the more likely it is that something can make it go *poof*.
On the flip side, though, the IIgs is made of pretty "chunky" circuitry, certainly by today's standards, a good chunk of which is 5v CMOS (which is a relatively stable process). I have some pretty serious doubts that a bit-flip that could cause a spurious interrupt is something that a typical IIgs owner would ever see in the lifetime of the computer. DRAM memory is the least stable part of the system so a soft RAM error would be far more likely, but still not the sort of thing you're likely to see every day unless you're living in a uranium mine. (In which case your "cosmic rays" will be from radon and other earthly sources. When radiation-induced soft errors were first observed in the 1970's the bulk of the problem was radioactive contaminants inside the chip casings decaying.)