DEV Community

Will Macpherson
Will Macpherson

Posted on Originally published at getantidote.com

Code review, automated tests and observed behaviour

Different checks answer different questions

A release decision can draw on several kinds of evidence. Code review examines a proposed change. Automated tests exercise selected expectations. Product observation follows a journey through the running experience. Each contributes a different view of the software.

Code review: inspect the change

Review helps a team discuss intent, implementation and maintainability. It can surface assumptions and identify paths that deserve closer examination. The question is what the change does and whether its design makes sense in context.

Automated tests: check chosen expectations

A test expresses an expected result under particular conditions. Its result is most useful when those conditions are understood. Passing a test tells the team something about that check; it should not silently stand in for every journey or environment.

Observed behaviour: follow the experience

Product observation asks what happens along a defined user journey. Does sign-in lead to the intended screen? Can onboarding reach its useful outcome? Does checkout open confirmation after an address change?

A finding becomes easier to investigate when it identifies the expected outcome, the observed result and the steps that led there. The scope belongs beside that evidence.

Bring the views together

Use the observation to inform the next investigation. A reproduced issue can guide a code review or become a useful regression check. A blocked step can reveal that the environment needs attention before the journey can continue.

Antidote is intended to complement existing engineering checks. The aim is clearer evidence for the people making the release decision, with unexplored areas still visible.


Originally published on the Antidote blog. I work at Antidote.

Top comments (0)