Hello again :)
Writing about Software Testing gives me so much fun, that I'm creating new articles much often than I expected at the beginning :) This time I will cover another famous big software crash. The last time I wrote about Therac-25 case (https://read.cash/@Otek/software-testing-when-computer-glitch-kills-race-condition-error-example-c5266bbf), today it's time to write something about Mariner 1 space program :)
Mariner 1 (R-1) - the first space probe of the Mariner program of the US space agency NASA. It was built as a result of cooperation between NASA (National Aeronautics and Space Administration), JPL (Jet Propulsion Laboratory), and USAF (United States Air Force). The plan for that mission was to send it using rocket Atlas Agena B to Venus to collect valuable data about that planet. That project was very ambitious, cause it was the first American planetary flyby of Venus.
Launch Day
Mariner 1 was launched on July 22, 1962 at 09:21:23 GMT from Cape Canaveral. Atlas-Agena lifted off from Pad 12 and at the beginning, the start seemed to be performing normally. But just after a few seconds rocket veered off course and continued to change direction in a strange way. To prevent a hitting a ground - after 294 seconds of flight, an order to self-destruct was sent to the rocket. So, just that after 294 worth 18,5 million dollars rocket + probe was destroyed. The question like always is why?
Two errors
In effect of long 'post mortem' investigation reports said that there were two errors that went to the disaster. What is interesting if the former had not appeared, the latter would not have happened either, and everything would have gone according to plan.
First error: Immediately after take-off, it turned out that the rocket antenna practically did not receive any signals, including navigation commands from the air traffic control center. However, the designers of the rocket were prepared for such an eventuality and created a plan B in advance of such cases. This plan ordered the onboard computer to cut off inaccurate signals from the antenna - completely ignore them and run the emergency software, which was to continue the flight according to a predetermined plan. So that has happened in that case, but instead of flight in a straight line - a violent cycle of rapid and significant flight course corrections began.
Second error: there was a tiny error in the equation in the emergency pilot program. The correct version is shown below.
The software in Mariner 1 lacked this little line at the top: "_". That mistake was made at the stage of handwriting one of the equations. The prescriber missed the dash in the equation. But why that one character is so important?
This dash indicating that the guidance computer should average (smooth) the data. Without it, all speed changes were considered serious by the navigation system and caused a sudden course correction. Sir Arthur Charles Clarke English science-fiction writer describe it later: "The most expensive hyphen in history.". As a result of that error and strange behave of rocket ground control send an auto-destruction signal to prevent Mariner to hit the ground.
Why does that mistake happen? Well like usual pressure and time were the reason. During that time cold-war was a really serious thing and CCCP and USA were in the middle of the space race. Everyone wants to be first - so politics rushed NASA to launch Mariner as soon it is possible. For that reason, there was not enough time for quality control.
What is most important lessons were learned from this failure and the planned mission was then carried out by the identical Mariner 2 probe (of course with fixed software), which was launched on August 27, 1962.
