In the age of tubes and transistors, how did the Russians or the USA soft land probes on the moon?
There is also a wealth of material on the US Surveyor program available through the NASA Technical Reports Server,
http://ntrs.nasa.gov/I've read detailed descriptions of the Surveyor flight systems: its star trackers, inertial reference, attitude control engines and yes, its electronics. It's a reasonable design for the mid 1960s, and being an unmanned robot it could accept a much greater chance of catastrophic failure than Apollo. Indeed, two of the seven Surveyors were lost during landing.
Although it had a landing radar to measure distance and velocity to the surface, Surveyor could not detect obstacles such as craters and boulders. Even the radar it did have could be confused. During the landing of Surveyor III, famous for its later visit by the Apollo 12 crew, its radar got confused and did not shut down the engines when it reached the surface. It bounced high off the surface twice before its engine was manually shut down by ground command and it settled down -- on the slope of a crater.
This is exactly the kind of thing that made a skilled human pilot so valuable; he could
see where he was going, avoid obstacles and pick a safe place to land, and no flight demonstrates that better than Apollo 11. Even today we don't really have much of a clue how to write computer programs that can analyze the output of a TV camera in the way that the human brain can interpret what it sees through its eyes.
The Lunar Reconnaissance Orbiter is returning photographs and maps of the moon that are so detailed that they can be used to pick safe landing spots for future missions, manned or unmanned. This was one of its major mission goals. This has never been done before; even with intensive Lunar Orbiter and Apollo orbital photography, the Apollo landing missions still had to depend on their pilots to avoid the obstacles that were simply too small to be seen from orbit at that time. This information will be especially of help for future robotic landers.
Even today spacecraft electronics are not nearly as advanced as you might think. They generally lag the technology on the ground by at least 10 years because of the necessarily ultra-conservative nature of spacecraft engineering. Commercial-grade parts are not necessarily excluded, but everything new has to be analyzed, tested and finally qualified for space before it can be used. The traditional design principles for spacecraft are still valid: a part that isn't present cannot fail, so KISS -- Keep It Simple Stupid. If you can still accomplish the mission while keeping some part or system on the ground instead of putting it in the spacecraft, then do so. If you haven't tested something, then it does
not work. Even if it seems to work, test, test, test, and test again.
And you can still be damned-if-you-do, damned-if-you-don't; I personally put much of the blame for the failure of the first Ariane V on overly strict rules on the selection of the computer processors for the guidance system (specifically the inertial platform).
Radiation susceptibility is always a big issue, especially for spacecraft designed to fly in or repeatedly throgh the Van Allen belts. Unlike manned missions that last no longer than 2 weeks, robotic spacecraft are expected to operate for years, accumulating considerably more radiation and almost certainly being exposed to a number of solar proton events. Here the older technologies actually have the advantage as they tend to be more radiation-resistant than newer, very highly integrated circuits with very small feature sizes.