My feeling is that both types of testing check system behavior for random sequence of data.
However, the difference is that:
- in property-based testing there are multiple equivalence classes; I define expected behavior and generate a set of random values for each class separately
- whereas in fuzz-testing there is only on equivalence class that the system does not break (i.e., throw exception or core dump) for any random value,
Is my understanding correct?