The video game industry requires increasing economic and human capital to develop impressive, complex works, capable of surprising an audience accustomed to the blockbusters. That comes at a price. Whether it is level design, 3D modeling or rendering of large environments, the data load is superlative; development times, too.
From the San Jorge University of Zaragoza, the research group Software Variability for Internet of Things (SVIT) he has endeavored to turn the situation around through the use of engineering. its content suggestion engine not only has he been able to reduce development time in some sections from months to hours thanks to the reuse of previous work, but has been recognized in the prestigious Humies awards. MeriStation has interviewed Carlos Cetina, Software Engineer at SVIT, to explain the operation of this suggestion engine, its direct applications in development and the possible uses it can offer to the industry in the future.
Time, that final boss impossible to stop … but to avoid
“The scientific community does not yet think that the video game is so in the focus of research,” he begins by telling this medium, making it clear that he is not a video game developer, but that both he and the rest of SVIT are researchers in the engineering of the software. For different reasons, the world of interactive entertainment entered their lives, as did other areas in which they work. The objective of your work is always the same: that the support of a computer can serve as a complement to human work.
“Video game development requires an amount of effort and hours on the part of the developers that, if you think about it, there are millions of artifacts to keep in mind. They are creating worlds. Video games have reached a scale that transcends what a developer can do”, He emphasizes. And that is why it is necessary to rely on technological engineering, to reduce human effort as much as possible until reaching the objectives thanks to computer support, thus streamlining processes that can take months.
Using technology as a support: a digital co-worker
During these last thirteen years they have been working in classical engineering; But software has become anywhere, even induction hobs in home appliances. Thanks to these investigations in the industry they have acquired knowledge that, now, becomes knowledge applied to the area of electronic entertainment; the one that concerns us. As soon as the Universidad San Jorge included in its academic catalog the possibility of studying video game development, the SVIT team came to the fore. His achievements are now recognized internationally, with Facebook interested in leveraging his suggestion engine.
So what do they do at SVIT? Cetina admits with laughter that it is very simple: “we are dedicated to accelerating software development.” For many years in the classical industry; now, applied to the video game sector.
How does your suggestion engine work? An innovative and useful idea
“The operation starts from the content,” explains Cetina. They always need ready-made examples, as the engine compares, parses, and creates new examples. “You can focus on objects, weapons, non-controllable characters (NPC), final bosses… There are many types of content. In our case, we focus on the final bosses. We use them as a starting point and the engine creates new final boss candidates. To do this, choose some of the starting criteria under specific selection criteria and cross or mutate them. Thus we obtain variants of the ones we have taken initially. How do we know if those variants are good or bad? With a criterion that scores them “, completes the award-winning member of SVIT. “To score them we run a simulation between a simulated player and a newly created variant.” Once they determine which are the valid examples (new final bosses), the copy is created and is ready to be used. “The engine is capable of replicating the vision of the creators”, He concludes in his explanation.
“Our engine works like biology”
“The idea arises from having contact with the videogame development people at the University; specifically the people of Kraken Studios, which are possibly a spearhead here in Spain. At first we spoke in very different languages, especially because we came from the academic side, the classical industry; but little by little you start talking to them and you realize that we are in the same place and it would be very convenient to transfer our knowledge ”.
When they learned about the problems that were transmitted to them, which were based mainly on the need to work for about five years on all those ideas they had in mind, they realized that the solution was to speed up processes. This is how SVIT began to work with them to identify what could be optimized and how they could complement it.
“Our engine works like biology. You have some content, in this case digitally, that you take it to a computer and you start to cross it, to mutate, to evolve, until we conduct all that evolution taking into account all the content that the developers have previously made ”. Its suggestion engine does not replace, but complements. “It’s a help,” he insists. “In the end, it’s like you end up having one more coworker, but digital”.
Thanks to this type of technology, Cetina gives as an example cases such as Cyberpunk 2077, by CD Projekt, in which what are now inaccessible buildings in the future have all their businesses and activities, since much of the work already carried out can be reuse, adapt and improve without having to start from scratch.
Facebook’s approach is just one of many they have had in recent times. One of the examples he gives us is Tequila Works, the renowned Spanish studio responsible for works such as Rime, The Sexy Brutale or GYLT.
The Humies awards, a bronze that tastes like gold
After so many years working in the classical industry, from SVIT they have matured as a research team to the point of presenting themselves with guarantees and confidence for the Humies awards, granted by ACM-SIGEVO, one of the most prestigious international evolutionary computing associations. Humies recognize those jobs in which a computer achieves better results than a human being. With their third place finish, they have only fallen behind the Google research team and Stansford University. Cetina hopes that in the near future more studios will be interested in its suggestion engine and, little by little, the video game industry will become a more sustainable place.