Most bugs occur in those areas of the system that developers changed since the last release. As testers, we thus attempt to test those areas especially carefully.

However, our research shows that substantial amounts of code changes are shipped into production untested, where they cause the majority of field bugs. And this happens even if teams have very sophisticated test planning processes.

The key reason for this is insufficient communication from developers to testers about what was changed. In our experience, information documented in tickets or user stories is typically not enough.

In this talk, I present test gap analysis, a novel approach that combines static and dynamic analyses to discover untested changes. I outline experiences application in practice from different domains, in international teams with up to 20 developers and 30 testers and describe how it helps to keep development and testing in sync. I also demonstrate best practices for hotfix-tests, continuous integration and test planning.

Slides