In computing, software testers and software engineers can use an oracle as a mechanism for determining whether a test has passed or failed. The use of oracles involves comparing the output(s) of the system under test, for a given test-case input, to the output(s) that the oracle determines that product should have. William E. Howden first defined and used the term in this sense in his article "Introduction to the Theory of Testing". Additional work on different kinds of oracles was explored by Elaine Weyuker.
Oracles often operate separately from the system under test. Method postconditions are commonly used as automated oracles in automated class testing. The oracle problem is often much harder than it seems, and involves solving problems related to controllability and observability.
Common oracles include: