UI automation – it’s not free
summary
Covers why Selenium WebDriver-driven UI automation still carries large hidden costs: one real project grows to 1,000+ UI tests running 22 hours/day across 100 servers and 400 agents with only a 70% pass rate, long debug cycles and heavy maintenance. Proposes shifting effort down the pyramid: add unit tests, add APIs (author uses SOAP UI, Maven, Jenkins, Rest Assured, SonarQube) and trust lower-level checks — example: same check as UI takes 4m but API takes 30s.
AI-written summary · about
chapters (14)
- 0:00Intro
- 1:13Why UI automation is popular?
- 3:38Why UI automation is not free?
- 4:43Baggage: Long execution times
- 8:01Lessons learned
- 9:23Baggage Maintenance
- 13:15What's Next?
- 19:57API or web services layer
- 24:28Testing web services with SOAP
- 27:04Web services testing
- 28:09Ul testing layer
- 32:16Writing good scenarios
- 35:47Exploratory Tests
- 38:39Conclusion
related talks
21mLowering End to End Testing Barriers
13mBeyond the hype: What test automation really costs
32mLiving on Crotches or UI Automation Antipatterns
56mThe End Game
38mFreeing up Selenium
51mRedefining test automation
05mTaming the Testing Beast
01h 08mTest Automation != Less Testers || Faster Testing || More Time For ET