In a simplified view, each time the code performs a data manipulation, the software sets out the various ways that computation could go wrong (divide by zero, overflow errors, and so on). It then ...