Come together
Unity Technologies adds games to its enviable list of achievements
There’s a common refrain that we hear repeatedly during our trip to the Copenhagen offices of Unity Technologies, the maker of one of the most popular game engines on the planet: “We don’t make games.” VP of engineering Brett Bibby says it a good few times, adding, “We are the 600-person-strong engine team for every developer”. Or, “We’re successful only if our customers are successful.” Smart stuff, you’d think, for a company in the business of making tools – but in an effort to beat the competition, Unity has now started developing its own games, too.
Since its foundation in 2004, Unity’s central mission statement has been “democratising development”. That means producing tools intended to make it easier for developers to create games – and offering them for free to smaller studios. A big part of that ambition has been about helping developers to fill game worlds with minimum effort. A quick glance at the Unity Asset Store will turn up free packages of highly detailed trees and rocks, or parts enough to build your own 2D platformer.
Unity’s tools for this are only getting better, as we’ll find out during the course of our visit – but sheer scale of content isn’t the only thing that matters for the company today. Fortnite, after all, has shown how you can have just a single map and still be the biggest game in the world.
Fortnite represents something of a problem for Unity. Along with other recent multiplayer hits including
Player unknown’s Battlegrounds and Ark: Survival Evolved, it was built on Unity’s main competitor, Unreal Engine 4. Moreover, its success has helped reshape the engine itself. Unity has described
Fortnite as a “development sandbox”, guiding the way TV-sized games get ported to mobile and pushing the engine to optimise its multiplayer performance for much larger groups of players. That makes Unreal even more attractive for anyone looking to develop the next world-dominating multiplayer game.
Meanwhile, developers using Unity – of which there are over a million active each month – are more likely to be creating smaller independent projects such as recent success stories Hollow Knight, Overcooked and BattleTech. On the horizon, the most high-profile titles being built with Unity include Moon Studios’ Ori And The Will Of The Wisps, and Campo Santo’s In The Valley Of Gods. All exciting games that will do well, but Fortnite has moved the goalposts somewhat. If Unity does have an equivalent, it’s Pokémon Go – but the workings of that game are so particular that it can’t really serve as a blueprint for success. “The next big first-party game for the PS5 being made in Unity,” Bibby says, daydreaming. “That’s what success looks like.”
In the meantime, though, Unity is planning to just produce its own paragons of what the engine has to offer today’s developers. So, yes, at least temporarily, Unity does make games. Not full ones, but sample games that it’s hoping will inspire developers. The first of which, probably not coincidentally, is a multiplayer FPS.
This as-yet- untitled shooter has been built by a team of six, formed last August specifically for this project. They’re led by programmer Peter Andreasen, who has spent around two decades on the other side of the fence, working on games such as Hitman: Blood Money and Kane & Lynch: Dead Men. “We wanted to do a multiplayer game, to show users how you can do it in Unity,” Andreasen tells us. “We heard a lot from our users that they would like to do multiplayer, but it’s a pretty tricky thing to enter into, so we’d like to give them a starting point.”
This is the entire point of the game’s existence. Everything from its asymmetrical design to the backstory of its world was picked because it serves the technology on which it is built. The action sees a race of alien terraformers struggle with native robotic miners, as each group tries to capture and defend three points on the map.
This set-up was, Andreasen admits, “an excuse to produce a wide range of content”. The two races allowed his team to develop two distinct character classes, with their own weapon types. The fiction, such as it is, led to a map that runs from barren wasteland, through industrial interiors, to exotic flora.
The game wasn’t primarily designed to be fun – although there’s much whooping and hollering as the rest of the dev team duke it out during our visit. Like a regular tech demo, it’s a showcase for what Unity is capable of. On top of that, though, it’s intended to be a learning tool – for developers that use the engine, certainly, but also for Unity itself.
“Ideally, users who are building their own game can look in here and they can say, ‘Okay, this lighting, I’d like something like that in my game’, and they can open it up and see how it’s done,” Andreasen says. “We left all the tools in there. It’s not just the runtime aspect of the game, it’s also the different editor tools that everyone has to build in order to make a production. So we left them here for people to take and maybe use in their own projects.”
There’s a hope that this sample game might also provide an access point for people looking to take their first steps from playing to creating. “Many of us actually entered development by modding, or programming stuff that had been opensourced,” says Andreasen of his team. “For a lot of people, tinkering and experimenting is a very effective way of learning things.”
Meanwhile, within Unity, the sample game is a way of testing new features in a controlled reality before sending them into the world, like an internal beta test. Or, to use the tech company parlance, ‘dogfooding’ – using your own products to better understand their problems and thus, potentially, the solutions.
“For a lot of people, tinkering and experimenting is a very effective way of learning things”
The team’s glass-fronted office is located (intentionally, we’re told) right in the midst of the R&D department, to encourage colleagues walking past to poke in their head and mention a new feature that could do with testing, or to ask for feedback on something.
It all sounds great in theory. But in practice, what have they actually learned from the project? Andreasen can’t point to any dramatic changes to Unity that have been specifically triggered by the development of this game. It’s mostly improvements within optimisation and performance, he says. Okay – so, given the team don’t have to worry about sales like most game developers, what does success look like?
“Because all of us have a background in game development, we constantly think about, ‘What would someone like me find useful, if I was on the outside?’ So the best test of whether something is useful is if it makes other people successful. If, in a couple of years, I meet some game developer who has made a runaway hit with Unity, sold ten million copies, and I figure out that team started with this sample game, even if they ended up changing every single line of code… I’ll be very happy. That’ll be success for me.”
If the sample
game is a reflection of what off-the-shelf engines need to offer developers today, we’re also shown a more traditional tech demo, with a much grander scope, that shows what Unity will be capable of in future. A flying car, looking like it could have been ripped straight from the original Blade Runner, darts between brutalist towers lit by dusty sunlight. It’s Gormenghast as cityscape, a sprawl of gnarled skyscrapers peppered with neon signs and exoskeletons of scaffolding, that stretches horizontally and vertically as far as you can see.
Flying through the midst of all this spectacle, it’s hard not to feel the influence of a certain E3 demo on this cyberpunk vista. But, honestly, it makes Night City look positively provincial, albeit with a few important caveats. This is a city populated only by other airborne vehicles, forming lines of traffic in the sky above and below, rather than the streetlevel bustle that CD Projekt has been showing off. And it is, at least for now, strictly a tech demo. Nonetheless, the sheer scale is dizzying. There are somewhere in the region of ten million game objects in this demo, we’re told, and 3.5 million of them are visible at any one time. A large-scale Unity game apparently contains around 100,000 – equivalent to just one building in this city.
Remarkably, this scene was all built by two artists, on loan from another project, in the space of three weeks. At least, its skeleton was – production of the individual high-res assets was outsourced to external artists. Unity has built an entire city just to show off the latest feature offered by its game engine: the not especially sexy-sounding ‘nested prefabs’. The team use various analogies to explain these – “building blocks”, “Russian dolls”, “Lego” – but ultimately, they’re just a smart extension of how Unity handles assets.
“A prefab is a reusable object,” Nikoline Hoegh, user experience designer at Unity, tells us. “An object could be a can, trash bags, a building – these are all game objects in the world. I can put an object somewhere, create a lot of copies of it, and if I update just one of them, they’re all going to change. This was already the existing system, and the way that games and worlds are built with Unity today. What’s new on top of that is the ability to basically take a prefab and put it inside another.”
This allows smaller objects, from walls to window shutters to air-conditioning units, to be assembled into bigger tile-sets, which can in turn be combined to form blocks as large as an entire skyscraper. It is, essentially, tile-sets all the way down. If this sounds like it would make for repetitive environments – well, that’s not visible in the final result. Risk crashing your hovercar to go in for a closer look, and you might notice some of the same neon signs dotted around, but the overall effect is of something bespoke. The buildings are formed, almost coral-like, out of apartment-sized protrusions, and it’s hard to spot any patterns in the way they’re arranged.
Unity claims this kind of environment simply couldn’t be built without nested prefabs. It believes the feature will solve a problem that’s putting strain on even the biggest developers as they push for increasingly open worlds – the sheer workload involved in creating high-fidelity environments at scale. “If you look at the explosion in resolution, I can’t build 50 levels of 8K-display-worthy rich photorealistic content,” says Bibby. “It’s too much work.”
There is talk of spinning out the Mega City project into its own sample game in future, but it’s just one of the options being discussed. “What Peter’s group is doing is moving us from being [neutral] Switzerland to something more opinionated,” says Bibby. “If we were going to build an FPS, this is how we would build it – and we want to do that in different genres.”
The idea is to package each sample game with all the bespoke solutions you’d need to make something in that mould. RTS, MMO and fighting games are all on Unity’s hitlist. Each of those is a primarily multiplayer genre, but each requires very different technology, from renderer to network architecture. This is part of a broader move by Unity, towards offering a more modular tool to developers. “We believe that one size does not fit all,” Bibby says. The idea is that it will offer specific engine packages for individual genres. And beyond that, developers will be able to customise Unity, stripping away any tools they don’t need in order to create the engine which best suits their specific project – whether that’s an eye-catching indie game or, perhaps, the next world-dominating multiplayer shooter.
There are somewhere in the region of ten million game objects in this demo, we’re told