"Automation Testing with Organised Chaos"

To help our regression suite to be robust, we typically make our automation tests deterministic. So every time the suite is run, it performs the same actions in the same order. This helps catch regression bugs but will not uncover new bugs.

Humans are generally less-deterministic, and have a better chance of uncovering new issues, but take more time. There are also “monkey tests” which inject random events and check if the system breaks, but these issues are hard to reproduce and make it difficult to zero in on the root cause.

How about automation that works somewhere between human and random testing, somewhere between deterministic and exploratory testing? A solution that exercises the system under test differently each run, while using deterministic test cases?

In this session Ashok evaluates these questions and show the ways that flow graphs and AI techniques like evolutionary algorithms can be used to enhance your automated testing. This approach is particularly useful for integration and end-to-end testing.