Automated Checks

A deliberate rename of Beck’s Automated: the suite should run and judge itself without a human: self-checking, machine-runnable, wired into CI. Fowler calls the result self-testing code; Meszaros counts the “Fully Automated Test” among his goals of test automation.

The rename concedes the context-driven point (Testing vs. Checking): what runs unattended is checking. This is a property of the checks in a suite, not a claim that testing (the exploratory, questioning human activity) has been automated.

Part of Test Desiderata.

Measuring: binary in practice. Either a single command runs the suite and returns a verdict, or it doesn’t. Watch for humans in the loop: manual setup steps and eyeball-verified output are the violations.

Sources

• Martin Fowler, SelfTestingCode: a suite that runs and judges itself.
• Gerard Meszaros, Goals of Test Automation: the Fully Automated Test.
• The context-driven case for the rename lives in Testing vs. Checking.