You gotta love a design that can be fixed remotely, that far away, so long after launch (not to mention their inflight SW modifications for the Galileo mission, etc).
Engineers have successfully corrected the memory on NASA's Voyager 2 spacecraft by resetting a computer bit that had flipped. Reset commands were beamed up to the spacecraft yesterday, Wed., May 19, and engineering data received today confirm that the reset was successful. The Voyager team will continue monitoring the engineering data, and if the bit remains reset, commands to switch to the science data mode will be beamed up to Voyager 2 on Sat., May 22. Receipt of science data would then resume on Sun., May 23.
I'm not familiar with the computers on the Voyagers (I did work for a company that provided slipring assemblies for it). But I did work with simple computers around that time that are probably fairly similar. The PDP-8 was kinda neat as a learning platform.
The instruction set was available in a pocket-sized card about like two playing cards taped together. I think it had about 32 possible basic instructions. If you analyzed the instructions in binary, it was obvious, for instance, that one particular bit turned on the twos-complement-addition logic unit. The machine was so simple that one engineer could understand every single thing about the computer, from the programming down thru the function at the gate level.
A machine like that is amenable to troubleshooting.
Yeah, there are definite advantages to the computers from those days (from right before I started playing with them, unfortunately, so I didn't get an opportunity to learn on them).
V'ger was a ripoff of Nomad, but had a happier ending.
"er-ROR?!"
From the field of combinatorial analysis, 32 CPU instructions taken 10 at a time yields 64,512,240* as the upper limit for the number of possible 10-line programs. Upper limit because not all instruction sequences are allowed. *For 16-line programs the value is 601,080,390, the peak value for 32 instructions (hmmmm...32...16...Nyquist?).
The modern computer industry has failed to learn the lesson that compact, easy-to-memorize, human-to-machine instruction sets still allow ample room for sophistication and complexity. The problem is that simple instruction sets and programming languages don't promote job security, and make it extremely difficult to maintain the unnecessary software complexity required for "social engineering" purposes (e.g., forced web connectivity, whether you want it or not).
God forbid that the peasants should be able to write their own software that does exactly what they want. Let them run the "approved" software, or none at all. Let the operation of an IPad be the peak of their creative powers, heh, heh, heh.... Useless eaters.
*[EDIT] I was looking at the wrong line in my CRC Math Tables. Corrected the C(32,10) value and added the C(32,16) value.
"er-ROR?! Must...sterilize...STERILIZE!"
