As the Principal Quality Assurance engineer at Spongecell, I was responsible for automated testing and created a suite of smoke tests for our Web application using the Selenium testing framework.

I started out using the Selenium IDE, but quickly progressed to using Selenium RC. I translated our existing tests (developed using Selenium IDE) into Ruby and started writing them by hand. I took over a Mac Mini that we had in the office and set it up with Parallels to cover IE6 & 7.

I developed the tests on my development machine and checked the suite into Subversion. On the testing machine, a cron job fired every hour, synced up the Subversion repository, and ran the suite of tests. It output the results to a webserver set up on that local machine, so that folks could go to a URL and see whether any tests were failing. The system also sent warning emails to our QA list when tests failed.

At that point, the QA team (me, and later one additional tester - Allison Wamback), would investigate the failure manually and write up a bug to communicate the problem to the engineer responsible for that area of the product.