At the Beyond Pass/Fail event sponsored by Thoughtworks in late January at Skills Matter, Chad Wathington talked about strategies to start with test automation on legacy projects. He advised against automating all manual tests because not all tests would be equally valuable or important.
As the first step, Chad advised automating smoke tests first. According to him this is low hanging fruit that gives the team immediate benefits, as such tests are executed often and on legacy systems they catch lots of problems often.
After the smoke tests, he suggested looking at expanding the automation pack in four potential areas:
- Strategic areas of the application (that need to be tested often)
- High risk areas of the application (that need to be tested thoroughly)
- Easy to automate, long to execute tests
- Tests with often repeated steps
As the third step, he advised against automating other types of manual tests but instead actively automating tests around new functionality.
Overall, this strategy is designed to get a team started with automation quickly, give them some short term benefits but also not disrupt the work on releasing new functionality too much.
|
|