Re: Not Me But...
> And how long did it take to get everything booted up after an orderly shutdown?
A few minutes at most, either to reload a 3705 or even to reboot an MVS or VM system. There were no Terabytes of disks to fsck, there were no Gigabytes of useless libraries to load, and remote data communications paths rarely had bandwidths greater than 9600bps. (You could monitor data trafffic by displaying it on a DataScope and watching it go by.) The size of everything was expressed in K or M. The machines were very slow but everything in the IPL path was hard coded assembler language purpose built to get things running. The mainframe OS and the 3705s were independently restarted as needed.
Also, the 3705 was anything but a "simple beast". It was a computer with a complex OS in its own right, supporting async (aka Start/Stop), bisync (aka BSC, for older 3270 terminals), and NCP for native SNA devices like newer 3270 clusters. Larger shops had multiple person groups of systems programmers who were dedicated to the 3705s (we had about 10 scattered across various data centers). We had very imaginative names for them: I, J, K, etc. Cabling was huge and unwieldy, and duct space was at a premium in many areas.
And yes, the 3705 load button could have been accidentally clunked by an IBM CE documentation cart, but those were big and heavy and rarely moved. Usually the CE would just go get what was needed as it was easier. I do not remember the castors on those carts being crap; if anything they were extra heavy duty like everything else IBM manufactured. More likely was the cart hit a bump in the raised floor tiles and veered off course. I could see that. CEs screwed up just like all of us other humans did.