[duped]

In my experience, manual QA is a symptom of bad engineering culture where engineers inherit large systems with poor test coverage. Automated testing is only hard when it isn't considered from the beginning of the project.

A lot of manual QA is bad testing too. If you have a grueling and linear process to go through to test your software each release... why isn't it automated? Is it doing something that fundamentally cannot be automated? If yes, you're probably wrong - almost everything can be automated.

Humans need to be put in the loop when requirements are changing and problems need to be solved. Designing comprehensive testing systems may be the role of good QA, but actually running the tests by hand is not a great place to be regularly.

[isbvhodnvemrwvn]

I agree with repeated tests, they are wonderful at catching regressions - however exploratory testing is still something valuable which many developers overlook - a lot of us are happy with clicking through happy path and maybe some most obvious error scenarios, but there is a lot of applications where you can paste several megabytes of text in some of the forms - something that a more thorough exploration testing session would likely catch.