A fake clock always says 9 in the morning. What kind of stand-in is it?

• A mock, because it checks calls
• A stub, because it returns a canned answer
• A copy, because it runs the real logic

A test replaces the database with something the test controls. What is that replacement called?

• A stand-in behind a seam
• A second copy of the database
• A faster database server

A mock records which calls it received so the test can check them.

Circle one:   True   False

You want to check that your code asked the database for user 42. Stub or mock?

• A mock, because the test must inspect the call
• A stub, because canned rows are enough
• Neither, because databases cannot be replaced

You test a function that reads the current time. Where do you put the seam and why there?

• Inside the time formatting code, to test each line
• Around the clock itself, so the test controls time
• Inside the test runner, so all tests share one clock

A refactoring renames two helper functions and five tests fail, though behaviour is unchanged. What broke?

• The stand-ins returned the wrong canned answers
• The seam was placed at the boundary
• The tests assert on internal calls instead of results

Ana replaces the network with a mock and then asserts only on the final message shown to the user. A teammate says the mock is pointless. Is it?

• Yes, because mocks must always assert on calls
• No, because the mock still removes the slow network
• Yes, because only stubs can replace a network

Two designs: the code builds its own database connection inside, or it receives one from outside. Which survives refactoring better, and why?

• Receiving it, because the seam sits at the boundary
• Building it inside, because there are fewer parts
• Both equally, because seams do not matter