Our QA engineer Povilas Svetovas on the state of play in QA today

QA has come a long way in recent years, but with issues around representation, automation and underappreciation, have things come far enough?

For many in the games industry, a job in quality assurance represents the first rung on the development ladder. One of such is our QA engineer Povilas Svetovas.

Leaving aside matters of personal taste, critical acclaim and popularity, and wearing only your QA hat, which game – other than one of our own – has impressed you the most over the last 12 months?

Xenoblade Chronicles 3 for Nintendo Switch because of its complexity and size. The different mechanics and functionalities are harmonic, and the quality of the games’ technical side only leaves a number of bugs that I barely noticed.

What’s been the biggest challenge that QA teams have had to face in recent years?

One of the biggest challenges QA has had to deal with in recent years was the necessity to keep up with a rapidly changing technological environment and still ensure an excellent quality of video games. This includes software compatibility with the increasing number of devices and platforms, new operating system complexity, and new software version requirements for security and privacy.

QA teams have always been frontline combatants when it comes to crunch. How far along are we from the c-word falling out of use?

Yes, it’s been a long time since QA teams have been associated with unavoidable crunch time, and every tester has had to deal with it, sooner or later. But it is possible to avoid crunch time or lower it to a bare minimum. There are three keys to that: test planning, test re-planning according to the situation while executing the test plan, and good communication with project supervisors/developers.

From my experience, the most frequent cause of testers dealing with crunch is not giving sufficient attention to test planning and thinking it’s not that important. Consequently, the testing schedule becomes abstract and lacks details/depth. Spending a few days delving into the analysis and planning is better. There are times when test planning is performed by a single QA Lead without including all the QA people responsible for it, especially when it is necessary to determine the scope of testing, test execution time, and/or the risks involved.

The project’s quality assurance and control are a part of the project’s management, so proper involvement of the people related to the project helps avoid miscommunication, surprises, and inconsistencies. Of course, unpredictable situations which lead to crunch time are unavoidable, but it is possible to lower the probability significantly.

Unionization has recently become a hot topic? How has that come about and do you foresee increased levels of unionization across the industry – in QA especially – as inevitable?

It is difficult to predict the future in this situation. Unionization may become inevitable because workers keep seeking better working conditions. It will depend on such factors as the IT industry‘s response to worker needs/demands forming trade unions in the regulated countries.

What is your favorite commercially-available tool that you discovered or made use of this year?

This year’s biggest discovery ties with an open-source tool – which we’re delighted to have stumbled upon – is Open CV framework. It has huge applicability to video game automatization, and we keep integrating it deeper into our work. Other favorites include Google Vision AI and ML-based model ‘keras-ocr’ for text recognition.

From commercially available tools that are ‘ready to use,’ we discovered ‘Browserstack,’ which allows us to check our projects with real devices through its cloud service.

How embedded are AI/automated tools in today’s QA environment and how have they made things better, or perhaps worse?

Without question, AI/automated tools make today’s QA work faster, more effective, and improve quality. It lets manual testers concentrate on creative work and more improvisational defect searches, leaving monotonous tasks for machines. Every QA department should invest in two testing types – Smoke and Regression testing as they both devour time and are very costly if done manually.

What other challenges do you see on the horizon for QA and how can teams and individuals best be prepared to meet them?

ML/AI integration in the testing process as it changes the QA work principles since the possibility of defect prediction appears. Various analyses help with better performance, as well as additional automatization possibilities. The integration of these technologies into daily testing changes the QA workflow. There is increased demand for faster software releases, and QA teams must find ways to ensure software quality while meeting tight deadlines.

To achieve that, QA will need to get so much deeper involved in Agile and DevOps processes. Recently, it has become necessary to adapt QA processes and tools available to achieve good results in this case. How do we prepare for that? Understanding that change is inevitable and meeting it with a positive attitude.

Are you optimistic about the future of QA? Why?

Software quality assurance is not going anywhere and will always remain a part of software development. Bad software QA is ultimately detrimental and costly to the final product, so the importance of QA in video games and other software fields will remain the same and probably even grow.

But new technology integration in testing changes the work principles of QA specialists, meaning their work changes from universal to more specialized.

It has been said for a long time that manual testing is impossible to replace. Still, the demand is decreasing even in video games since more and more QA processes are being automated, and AI and ML are being used more often to predict possible faults in QA. To sum it up, a bright QA future awaits those who can adapt, are innovative, and study a lot