Changing ze software
esa-aardvaark wrote "someone will correct me I'm sure".
Not really, old colleague, but I'd say the failure was caused by using segments of software from an earlier ARIANE in a later model where it wasn't appropriate.
And here's another one: when commissioning a new spacecraft in orbit, instead of action 'A' happening when command 'A' was transmitted, action 'B happened when 'A' was transmitted and action 'A' happened when 'B' was transmitted.
Representatives of the French constructor of the spacecraft who were present were asked to explain. Pulling out an old envelope and consulting notes written on the back, the French engineer squinted at the scruffy paper and said "we 'av changed ze software just before launch..."
Not a rumour, I was present when it happened.