A Survey and Discussion of Automated Testing

Created and summarized by Brian Marick.

By automated testing, I mean automated execution of tests that have already been designed. The purpose of automated testing is to make tests cheaper to run than if you executed them manually. Most often, the idea is to rerun tests that have been run before: so-called regression tests. However, sometimes automation is necessary to allow a test to be run at all. Think of a test that tries to see whether a product can survive 1000 simultaneous users - that's only practically tested with an automated tool.

What's hard about automated testing is that the product changes, which tends to break the tests. If you spend all your time fixing broken tests, was it really worthwhile to automate them?

Key issues in automated testing are:

• deciding which tests are worth automating.
• finding and implementing an automation approach that insulates tests against change.

I wrote a brief and whimsical history of test automation for Software Testing and Quality Engineering Magazine. It contains a number of annotated links to articles on the web that I think are well worth reading. It also contains pointers to books. My favorite book is Software Test Automation by Fewster and Graham.

Bret Pettichord's Software Testing Hotlist does a good job of keeping up to date with new links.

Related Testing Craft Pages

• There's more about test automation in the index of techniques.
• Using Python to test a product with a GUI through a programmatic interface. (not done)

Be the first person to add a comment in the Wiki Forum at page AutomatedTestingSurvey.
(The Forum is explained in its FrontPage.)

Summarized Discussion

In this spot, the author of this page will occasionally summarize the discussion in the Forum.