Last week, a PC in my wife's medical office wouldn't boot, with the dreaded "no drive detected" error message. Figuring "what do we have to lose", I turned it off cracked the case, started whacking the C: drive with a screwdriver handle, and punched the power.<p>It booted.<p>The office manager who was watching me had to pick her jaw up off the floor. We proceeded to copy everything possibly important off it onto a USB drive, knowing that may well have been its last spin-up ever and it's next power off may be the final one.
That reminds me of the story about the magic switch: <a href="http://catb.org/jargon/html/magic-story.html" rel="nofollow">http://catb.org/jargon/html/magic-story.html</a>
Great story though you could have just pulled out the switch and soldered the wires together :-)<p>I look forward to the practical soldering test for developers at FANG's - every one on here can solder right.
Office 97 actually gives me a little bit of nostalgia. It's the last version for me that actually felt pleasant to use (assuming you turn off Clippy, which isn't much of a barrier).
The fun part is that there is def a disconnect between the dev team and hardware itself (or help desk). To resolve the issue, you only needed to short two jumpers on the motherboard to get it to turn on.<p>If it was a NON-ACPI hardware, then the hardware switch was wired into the power supply itself, which, then again you could've undone the cable and connect the circuit to boot the computer.
In high school, circa '97, I was working in a computer lab. As I entered the room, I was rapidly admonished "<i>don't touch the table!</i>" So naturally, I reached out, tenderly as possible, brushed the table with my fingertip. Sure enough, the computer immediately reboots. He cusses a blue streak, as he's been trying to install Windows but folks kept touching the table! I regret not talking him through the issue, because obviously nobody would use such a flaky machine... but I'd (quite reasonbly) lost his goodwill in my initial mischeviousness. To this day I don't know how the machine was so sensitive to, presumably, a miniscule change in capacitance. There's a possibly that it wasn't grounded and I was carrying a nontrivial electrostatic charge, but the table surface was insulating!
Turning off the computer or putting it in suspend mode is somewhat expected nowadays unless it is a server in a rack that runs untold virtual machines that get 'spun up' as needed.<p>Back in 1997 the idea of turning of an specialist idle PC that consumed company electricity was not the done thing to do. If you were a lowly office worker using Office 97 or its ancestors then you would turn off your machine at the end of the day and start the day with a ten minute boot time. It took a long time to get these efficiencies right and a silly amount of time was wasted. CRT monitors ruled the roost back then too, after 6 you would see lots of silly Microsoft OpenGL screensavers running.<p>I wonder what the actual power consumption is these days compared to then? It could be a factor of ten.
I'm surprised there weren't backups they could've restored? If the difficulty was setting up the build environment correctly, then even a duplicate hard drive for each build server would've saved them from a pinch.
The amusing but strange sub-story here is the build engineer claiming that the product may never be built. That is a type of endemic first class whining the type of which I am sick, sick, sick of.