A well-known piece of space flight lore is the tiny - by today's standards - capacity of the computer that landed the first men on the Moon. But perhaps just as notable is the fact that, primitive though it was, it was smart enough to avoid crashing even as it was being overloaded with data.
The story of the Apollo guidance computer has been dusted off amid coverage of the 40th anniversary of Neil Armstrong taking humankind's first steps on the Moon. In a phenomenon familiar to millions around the world, when the Apollo 11 lunar module, Eagle, was descending at about 10,000m above the Moon, its computer started flashing an error message.
Fortunately for Armstrong and fellow-astronaut Buzz Aldrin, rather than cause the system to freeze - infuriatingly common on earthbound computers - the 1202 error merely resulted in brief consternation at mission control in Houston.
In Houston, astronaut Charlie Duke was capcom, or capsule communicator. Only the capcom spoke to the astronauts in space. "When I heard Neil say '1202' for the first time, I tell you, my heart hit the floor," Duke said.
But reassurance quickly came from the computer specialist in the control centre that the error was not critical and the descent could continue.
Then, at 610m from touchdown, Aldrin reported another error, this time coded 1201. After further consultation, Duke radioed back to the lunar module: "1201 alarm. We're go. Same type. We're go."
The drama is recounted in Tracking Apollo to the Moon, the 2001 book by space programme member Hamish Lindsay.
To add to the computer error excitement, when they were almost on the ground and out of fuel, Armstrong noticed they were heading for a rocky crater. He manoeuvred the lunar module sharply off to the side. Once down, he radioed to Houston: "The guys who said we wouldn't know precisely where we are are the winners today."
"No sweat," came Duke's response. "We'll figure it out."
Where exactly on the Moon they were was one puzzle, but the more pressing one was why had the guidance computer been acting up? Historian David Mindell, whose book Digital Apollo documents the Apollo programme's use of computers, writes that as soon as Armstrong and Aldrin had landed, engineers in the MIT instrumentation lab, where the computer was developed, tried to duplicate the problem.
They worked out that the computer was being swamped with radar data from a system that should have been turned off during descent. Clever software design meant that rather than lock up the computer, it triggered an alarm, but carried on with the essential task of controlling the descent.
From simulated landings done in the lead-up to Apollo 11's launch, mission control computer specialist Jack Garman recognised the 1201 and 1202 errors as non-critical - he had made a note of them on a sheet of paper - so when the alarm was sounded he was able to signal that the computer would continue to function.
According to Mindell, Armstrong's successful landing despite the computer alarm was portrayed in the press as the triumph of man over machine, causing resentment among the system engineers. In his view, the only effect of the alarm was to distract the astronauts.
One astronaut in particular, Dave Scott, played a part in developing the guidance computer, and in 1982 talked about his experience of using it. Scott, who went into space three times, on the Gemini 8 (with Armstrong) and Apollo 9 and 15 missions, said it was so reliable they never had a backup.
"We had some glitches here and there but, to my knowledge, in the 10 years I spent with it there was never a real computer failure."
Even so, when it was his turn to land on the Moon in 1971, he wasn't going to let the computer take the glory. "I like computers and I believe in computers but it ain't going to land me on the Moon. I'm going to do that. If something gets screwed up then it is going to be me, it isn't going to be the computer."
The Apollo guidance computer was so successful that it was used in the first fly-by-wire fighter jet in the 1970s, paving the way for use of the technology in passenger planes. Rather that than Windows.
The Apollo programme's computers were among the first devices to use integrated circuits - or chips - which were developed in the late 1950s and are now found in almost every type of electronic equipment.