Terminology is a problem here. A crashed harddisk is clearly an exceptional circumstance. More specific terms is needed to distinguish errors in the code (eg divide by zero) from unpreventable errors like network failure.
It is clearly not. It is very much something to anticipate. At sufficient scale, it is guaranteed that it will happen. There is nothing exceptional about it.
> More specific terms is needed to distinguish errors in the code (eg divide by zero) from unpreventable errors like network failure.
Luckily we have such terminology already: Exceptions (caused by mistakes in the code) and errors (caused by external faults).
However, outside of the Java world, which flips the terms around for some reason, this seems to be the prevailing usage. It is also the only usage that makes sense based on what seems to be the prevailing understanding of what "error" and "exceptional" mean. Java's "exceptional conditions" being the ones that you expect to happen regularly doesn't make sense, but was presumably an accident in naming that just so managed to stick.
Of course, confusion is compounded by some language idioms using exception handling constructs to handle errors.