Earlier this month in San Diego at Sandbox Symposium, a co-located two-day games analysis event running before SIGGRAPH, a panel of EA execs discussed a new pre-production process their studios have had in place to aid games design via rapid prototyping.
The process and concept of rapid prototyping itself isn't exactly new, but EA's creation of a formal process, structured to enable individual developers to express themselves creatively much quicker, makes for interesting reading - especially given that the process has had direct relevance to a number of new and upcoming titles from Electronic Arts.
According to EA, its new internal methodology for quick preproduction of ideas and concepts means ideas can be refined quicker - and cheaper. It also raises the quality of the ideas and helps keep the vision of a game strong throughout its full production cycle.
In this two-part article, we've outlined the key details from each speaker during the panel talk – this first part details work done by EA’s Glenn Entis, Chaim Gingold (Maxis) and Jeremy Townsend (Tiburon).
Introduction to Rapid Prototyping
(Glenn Entis - EA's chief visual and technical officer)
"We needed to quickly change habits," explained EA's chief visual and technical officer Glenn Entis as he introduced the rapid prototyping methodology that EA has implemented and is trying to encourage amongst its teams.
He explained that in 2005 EA execs realised they needed to readdress how the company went about creating games from their very earliest concepts - the benefit being better ideas that had been more thoroughly tested, and cheaper production processes thanks to the watertight ideas devised.
That year saw workshops conducted at every EA studio to put this into practice. Each was two days long featuring four to eight teams - Entis described them as fast-paced, with exercises just 15 to 20 minutes long to give people just a short amount of time to test their problem solving skills with tasks like "build a physical prototype in 20 minutes". The effect, he said, was a preproduction process that "asked the questions that don't usually get asked".
"We would find very often that people knew they should be doing smaller prototypes, but that if they were in such a big company making a toy would seem too light" said Entis, explaining that perceptions of how EA staff thought they should behave were slowly broken down by this approach. "They wanted to be a bit more looser and more creative than they were, but were unsure how it would be seen in the place that they work."
He explained that designers and developers eventually, through this process, "learnt by doing". Designers reenacted, physically, how characters or in-game objects (such as an exploding truck in Burnout) would behave. The workshops mixed the construction of paper prototypes and improv theatre - an approach that at first jarred with EA's corporate, perhaps stuffy, image.
Each exercise would build on the on work done in the work shop, leading from a concept to devising feature lists, a creative brief - even culminating in what the core 'X video' would be like. Also, the local studio leaders co-lead each workshop, so Will Wright managed the Maxis workshops, and so on - meanwhile, everything was ultimately relatable to real games, so titles such as Medal of Honor, The Godfather and Burnout all benefited from the more focused way to refine and purify their gameplay and design ideas.
Said Entis: "People say you can't manage creativity, but you can identify all the barriers that stop creativity and find the things that encourage people to be more creative or enjoy doing it. You aren't waving a magic wand and telling to people to be creative - but you if you do those things you can manage the environment to encourage creativity."
After Entis, five EA project leads took to the stage to discuss in detail how the quick prototyping process aided their games…
Prototyping on Spore
(Chaim Gingold - Game Designer, Maxis)
Gingold opened by saying that current game design was like an aircraft carrier at sea,, "and the problem with such a big thing like that is that we end up with genre conventions designed to maintain a low risk as games are so hard to make."
So, he said, "What if we made something that was small and quick?"
By example, he showed a four-your old application he coded during a plane flight for Spore which looked at letting players place buildings in a landscape and letting roads grow around it. "It was really cheap," he said, and "I didn't care that parts of it sucked as I wasn't attached to it." Eventually, after just a few hours, he realised that the idea was no good - but the rapid prototyping of the idea was excellent.
He outlined three specific paths that rapid prototyping takes designers down.
The great thing, said Gingold, was that quality doesn't matter in prototypes: "If you come back with something good people will be pleased," he said, but by that same regard, "if you go off for a week and come back with something lame no one will be mad at you." The point being, that at least you tried.
Secondly, "you are free to fail as much as you want. When you make these softwares you want to seek out these failures - failure is good," said Gingold, explaining that you want to see what people think, and iterating ideas to find out which ones work and which ones don't mean that players can impact design early on when the idea is in its early stages.
The third point moved on from that, pointing out that "prototypes become very adaptable" - in other words, its easier to tweak and refine an idea in its early stages rather than when it's in full production.
There were three direct payoff, from those paths, he said, which benefits any designer: specifically, experimentation, learning and adaptability.
Plus, he said, this approach can work on any game, large or small.
"If you want to make something really big," he said, "you take your big problem and break it down into small pieces, and then put them back together.
"Spore is so ridiculously complicated - when the EA CTO saw how big the project was he wanted to cry." To manage the process and instill confidence that the grand idea itself would work, the team broke down all its elements to refine each component of the design.
The same thing is clear in something like Mario, asserted Gingold, pointing out that its clear that Nintendo has structured its production around core components like level design, Mario's movement animations, the game controls, 3D camera, and enemies. Said Gingold: "All these things are related but ideally you want to cut them apart to research them properly."
Rapid Prototyping with XNA
(Jeremy Townsend - Producer, EA Tiburon)
Given that EA's Tiburon studio is best known for its work on the Madden series of games, has recently put out the best-not-talked-about Superman Returns game and has since been given Tiger Woods PGA Tour Golf to look after, you'd think the studio was busy enough with franchises and sequels to worry about. But the studio also has a new IP team investigating new ideas.
Jeremy Townsend is part of that group - and as part of the wider move towards rapid prototyping and more thorough preproduction he talked about using XNA for bringing ideas to live quickly. The advantages, he said, were clear - it was in the easily executable managed code C# and handled multi-threading, plus it lets even programming-minded consumers have go. Of XNA Game Studio Express, he said: "Considering the price of an actual Xbox 360 development kit runs in the thousands it's a nice concession to the community."
He stopped short of completely sounding like a Microsoft press release, however, adding: "'I’m not sure XNA is going to be the Youtube for games as there's a very high barrier. 3D math is more horrifying than marriage for some people."
As for how developers could directly use XNA for prototyping, he stuck with the 3D issue.
"Stay away from 3D prototyping if at all possible. Most game problems can be solved in 2D, even on paper," he said. "The Play's the thing - think of 3D prototyping as a big gun, you only want to use it as a last resort."
Like Gingold, he said prototyping was a quick path to get to the gameplay, so "if something doesn't work you can correct away from it". He also echoed Gingold's sentiments on cutting and idea down into smaller pieces when prototyping as it will give you a clearer idea of how a game's various mechanics interact.
Townsend also said that prototyping was a great way for safeguarding your ideas for potential future developments. The .NET library for the Wiimote integrated by the XNA community meant that EA could be prepared for the inevitability of Microsoft and Sony releasing their own versions of the motion-sensitive game controller by making plenty of PC-based concepts.
The result has been supporting Townsend's investigations into new IP for Tiburon to develop - although the projects are currently unannounced (and may never be), he explained that mixing XNA with the Wiimote has meant the studio has created demos for the device in just a day and a half.
Check develop-online.net tomorrow for the second part of this report which looks at work done at EA Canada (both Vancouver and Montreal) and EA Tiburon on more new IP titles using the likes of Flash, Virtools and Gamesketching.