Part of the Dare Mighty Things series — Apollo | Estimated read time: 6 minutes
Jack Garman was 24 years old, and the fate of the Apollo 11 landing was now his to decide.
It was July 20, 1969. Seven minutes and thirty seconds before the Eagle was scheduled to touch down on the Moon, an alarm had flared on the display panel inside the lunar module. Armstrong's voice came over the radio — calm, but with what the transcript would later describe as "the slightest touch of urgency." He read out the number.
1202.
In the Mission Control support room where Garman sat, guidance officer Steve Bales was already calling him. "Jack, what the hell is going on with those program alarms?" The telemetry delay between the Moon and Earth meant that Bales and Garman saw the alarm about three to four seconds after it appeared before Armstrong and Aldrin. They had perhaps that long to respond before the crew — hearing silence — might take matters into their own hands.
Garman looked down at the sheet of paper under the plexiglass on his desk. He had written it himself, by hand, weeks earlier. Every computer alarm code that could possibly occur during a lunar landing. What each one meant. What to do.
He found 1202. He looked up. He called it.
"We're go on that alarm."
The Machine That Got Them There
In 1969, most computers filled entire rooms. The AGC had to fit inside a spacecraft. It occupied roughly a cubic foot of space. It weighed 70 pounds. It operated on 4 kilobytes of erasable memory — roughly one-ten-thousandth of what a modern smartwatch carries. Its fixed memory was hand-woven: skilled technicians at Raytheon spent weeks threading thin copper wire through tiny ferrite rings, each thread representing a binary digit. A single memory module cost about $15,000 in 1969 dollars. The AGC used 36 of them.
The software was written at MIT's Instrumentation Laboratory. They had to invent not just the programs but the entire discipline of writing software for a machine that would navigate humans through the vacuum of space with no possibility of a patch, an update, or a reboot.
One of those safeguards was priority scheduling. If the computer became overloaded, it would shed the lowest-priority tasks and protect the highest ones — navigation, guidance, engine control. When it dropped tasks this way, it would restart, clear its queue, and flag the event with an alarm code.
The 1202 meant: "I have run out of memory space for scheduled jobs. I am shedding lower-priority tasks. My critical functions are intact and running."
This was a feature rather than a bug. The question — the one that had to be answered in four seconds — was whether the overload would remain manageable or cascade into something the computer couldn't recover from.
Why It Was Happening
The Eagle's rendezvous radar had been left running in AUTO mode during the descent. In AUTO mode, the radar was generating rapid, spurious signals the computer was trying to process — a steady stream of data it didn't actually need, flooding its scheduler like someone repeatedly hitting a request button on a website that's already under load.
Normally, the AGC ran at about 83 percent of capacity during powered descent. The radar was adding another 13 percent. When Aldrin entered a command asking the computer to calculate altitude data, it pushed the system over the edge by just enough.
The computer ran out of processing time, threw the 1202, shed its lower-priority tasks, restarted, and kept flying the spacecraft.
The List
Several weeks before launch, a simulation exercise had almost derailed everything. During a practice landing run, the 1202 alarm appeared. Steve Bales called an abort. The simulation controllers declared it a bad call.
Gene Kranz sat Garman down afterward and told him what he wanted: every single possible alarm code, including ones that had never been seen in simulation, with the correct response beside each one.
Garman spent a week on it. He consulted the MIT engineers. He cross-referenced the flight rules. He wrote his answers on a single sheet of paper in his own handwriting. Bales had a copy taped to his console. Garman kept his under the plexiglass.
Neither of them expected ever to use it.
Five Alarms
The first 1202 fired at seven minutes and thirty seconds before landing. Garman called go. The descent continued.
Fifty-four seconds later: another 1202. Garman called go again.
Two minutes and twenty-four seconds later: a 1201 — a variant of the same executive overflow family. Same cause. Same answer. Go.
Two more 1202s followed in quick succession, the last firing less than three minutes before landing when the Eagle was less than 500 meters above the surface.
Each time, Garman checked the list. Each time the answer was the same. Each time the call went up: go.
Armstrong, watching the surface approach through his window, saw the boulder field below and took manual control. He found a flat patch beyond the crater and put the Eagle down with somewhere between seventeen and thirty seconds of fuel remaining.
"Houston, Tranquility Base here. The Eagle has landed."
What Would Have Happened
If Garman had called abort on the first 1202, the mission would have ended there. Armstrong and Aldrin had not been briefed on program alarms. They were entirely dependent on the people in Mission Control — who were entirely dependent on the one engineer who had spent a week nobody asked him to spend writing down every alarm code he could find.
"You don't realize until years later," Garman said afterward, "how doing the wrong thing at the right time could have changed history."
Garman went on to spend thirty years at NASA, eventually serving as Chief Information Officer at Johnson Space Center. He retired in 2000. He died in September 2016.
He was buried without most people knowing his name.
The list is still out there. A sheet of paper, handwritten by a 24-year-old engineer, contained the correct answer to every alarm that could have stopped humanity from reaching the Moon.
This article is part of the Dare Mighty Things series — Apollo branch.