by Ondřej Gróf
One of the significant advantages of test automation is running tests on a vast amount of configurations at the same time. With the growing number of configurations, the probability of test failure is increasing. This problem causes that QA engineers lose a lot of time by tedious checking all unstable configuration and comparing them. What if a solution exists which saves their time?
We were aware of how much time QA engineer spends on that, and we started to look at the solution which transforms this problem to the most simple solution as possible. We created an algorithm which processes all test results from all test combinations and generates a clear chart where the failed test configurations are clustering by the common core of problems. Thanks to the clusters we can identify how many issues our program has as quickly as possible.
See how to use data that you have already generated to stop doing tedious routines. We will share with you how our story of using math in QA at Avast began
Ondrej has been a member of the QA Automation team at Avast for three years. With a team, he develops a test framework, cares about the test infrastructure and helps QA engineers to make a test result data as readable and understandable as possible.
He follows the latest trends in QA like continuous testing. In this case, he is aware of how much data is generated from automated tests and how important they are for an efficient development cycle.
Where you see Software Testing in future (in next 5 to 10 years)?
The testing in 5 years will redefine the description of Tester/QA engineer role. The tester will have to know how ML, AI and data analysis work. I believe that AI will generate almost all regression tests only from data of user’s behavior. However today’s testers will still find a job but only for specific projects.
Is testing long-term dream job forever or just on step/stop on your life journey?
My dream job is the one which is not about daily routines. Every day I want to explore or use new things and software testing exactly fit into this definition so I could do it forever.
What is the biggest obstacle or roadblock in SW testing that is slowing down improvement and innovation?
The people mindset that standards and routines are enough to check the quality of the project. Testers and QA managers should do things and tasks more experimentally.