Automate your way to ineffective testing with my twelve step program…
How to implement a company-wide test automation framework effort and fail (usually).
1. Choose your test automation tools without ever thinking about what problem you are trying to solve, or what tests are important to automate.
2. Choose the tool that will look best on your resume. Refuse to acknowledge the test results from any tool which would not increase your market value.
3. Have non-testers and people with no automation skill choose a tool and mandate its use for the entire organisation, without regard for the differing needs of diverse teams.
4. Have your test automators duplicate all of the automation that your development team is already performing in another tool.
5. Hire in skilled test automators to build automation suites, then have them hand over all of their automation code to a test team with no automation skills.
6. Ensure your skilled automators never get to work with each other, and build their frameworks in totally different ways.
7. Ensure that all of your test automators (who are unaware of each other) write tests to cover the functions that other automators are working on.
8. Make sure that there is no easily accessible central place for automation code to live.
9. Avoid creating and communicating standards for versioning, coding and merging of changes to shared automation code.
10. Avoid talking about test strategy. Ever.
11. Only execute tests that your tool can easily perform.
12. Have your regular test team automate every test they can think of.
ha ha…. Nice one…. I am learning QTP now.. Why? Cause our US people use it… It seems ok but I’m not sure it can do everything properly…. But at least Mercury get ****loads of money out fo us….
Indeed they do. I’ve just finished at an organisation where the test team is a little confused. On one side, the hired-gun test automators are telling them that they need to learn how to program. On the other side, their managers believing the tool-vendor marketing and are telling their testers that they don’t need to learn how to program.
If a team’s automation problem is simple enough to be solved using record-playback, a simple macro-recorder (or Selenium IDE) is probably going to do the job, and save them a bucket of money