We should really look at the classic failures at software.
So "Ariane 5" is one of the earlier rocket launchers. On the mid-flight, within a few minutes - or within a few seconds rather - of taking off, the launcher exploded. Now the reasons probably one could look to. The case study indicates the acceleration was a little bit more in Ariane 5 compare to the Ariane 4.
The other classical example is a Therac 25. It is an accelerator. It is a project developed by French and an American together. The Therac 25 was the system, is basically an accelerator utilised for the treatment of cancer. Around 1987 about six patients died while they were given this radius-treatment. This Therac 25, the software was used on Therac 6, Therac 10, Therac 20, this was utilised for a long period, though the Therac 25 is more automatic system, but it basically used the same software as the earlier ones. And the system failed.
A very similar thing if we look at the failures in the bridges that happen all the time. These were constructed with the help of the software.
We do not really know what has happened in Chernobyl. But probably, if there is a good software, it could have prevented. Though it is not the software fall that has caused the problem, but a good software could save the situation.