In software testing, how are error, fault and failure related to each other?

Explanation

In the dependability taxonomy used by IEEE and by Avizienis et al., a fault is the defect lying dormant in the code, an error is the incorrect internal state that arises when that fault is activated during execution, and a failure is the externally visible deviation from the required service. The causal chain is therefore fault → error → failure, which is option 4. Options 1 and 2 are wrong because they break the chain and claim the third term is unrelated. Option 3 states the alternative textbook chain (human mistake → fault → failure), which uses "error" in the sense of a developer's mistake rather than an erroneous program state.