Epic advice from epic studios, from a Epic designer "Lee Perry" -

"Don't be the kind of designer that writes a document then wanders around the street like a homeless guy, trying to get someone to make your game ... Don't be so dependent on everyone else to make your stuff happen." 

November 1, 2011
Lee Perry, lead designer at Gears of War developer Epic Games, encouraged game designers to take their fate into their own hands, saying they should use tools to prototype and "sell" interesting ideas to the rest of the development team. By using proof-of-concept maps instead of written design documents, designers can save time and more clearly show how a concept will work, in video form, he told an audience at the Gamasutra-attended Montreal International Game Summit on Tuesday. For example, one day Perry was sitting at his desk using Epic's Unreal Engine 3 scripting tool, Kismet. Playing with a prototype when developing the original Gears of War from 2006, he increased the size of a character model by 25 percent, then added a helmet to the creature's belly to give it a big gut. He added a "Booooom" sound, showed it to others in the studio who loved the idea, and the iconic Gears "Boomer" enemy was born, in prototype form. Relying on a proof-of-concept rather than a written design doc (basically, proof-of-concepts are visual design documents, Perry said), means that designers can spend fewer resources and time to try out a new idea. It also is much clearer whether or not that idea will work. Written documents have a place, but more for explaining ideas that were proven through proof of concepts. This reliance on designers was evident for the sequel, Gears of War 2. Perry said Gears 2 only had one programmer working on it for the first six months of development. Perry stressed the importance of putting tools in the hands of competent designers. Other instances of prototyping using Kismet included the creation of the Bloodmount, giant "cover worm" and Shield Boomer enemies, all of which shipped in Gears games. One of the cover elements in the Gears series allows players to kick a table or other object over and use it as cover. That feature was a Kismet effort that worked so well, Perry said Epic just left the Kismet version in the game, with no additional coding. "Yay, free feature!" he joked. Another upside of this method of prototyping isn't just discovering what's usable, but also what's not. Epic tried different camera systems, weapons and enemies that just didn't work. But with the rapid pace of the prototyping, scrapping an idea might mean only losing maybe a 8-10 hours of work. Designers implemented existing assets in resourceful ways. For instance, in Gears of War 3, players can step into and use a powered exoskeleton called a Silverback. An invisible Boomer served as the basis for the suit. When you cut through the arteries of the giant worm in Gears 2, those cutting points are invisible Wretches (ape-like Locust enemies). Typically, after everyone gets on board with the prototype, then the idea goes through the regular process of polishing and placement into a game. Creating a proof-of-concept map is best done by as few designers as possible - one is optimal, said Perry. "All this means is it's a person, maybe two, but keep it as few people possible to keep this pure, fast vision," said Perry. "...All that these are supposed to be doing is create interest within the team." It's the "salesmanship" aspect that is also a benefit to creating proof-of-concepts, as it lets designers champion their ideas. At one point during the development of Gears 3, Perry created controls for the enemy beasts - while there were naysayers (the wide array of enemies moved and attacked in various ways), a quick proof-of-concept showed that controlling these beasts worked, and it got a rise out of the team. So then the well-received multiplayer Beast Mode was born, and shipped in Gears 3. Creating these prototypes is not as complicated as it might seem, he argued. Perry said designers can repurpose trigger events, move objects, attach objects and triggers, play sounds and visual effects, use cameras like any object, dummy fire weapons or trap controller inputs in order to create something new - no coding required. "With any two or three of these things, you can prototype a lot of cool stuff. It's not as complicated as it is in your head," he said. "...Just because [a designer is] not a programmer doesn't mean that they cannot make cool stuff that you can learn from." Likewise, a coder can be more than the clean-up crew, and try their hand at designing using this method, he said. "Don't be the kind of designer that writes a document then wanders around the street like a homeless guy, trying to get someone to make your game ... Don't be so dependent on everyone else to make your stuff happen." He said that designers have the tendency to become the food critic instead of the chef -- that they become naysayers instead of the doers. While the food critic approach might work for some, Perry said, "I believe designers need to be like chefs. They need know how to cook. ... You need to be able to criticize, but you need to be able to know how to cook." Perry told attendees to find people on their teams capable of this kind of prototyping, give them tools to do this and have them use assets from unexpected places. And most of all, don't let naysayers talk you out of an idea until you've proven whether or not that idea can work. "If there's one thing you take out of this, find a way to take lots and lots of small risks. ... The project will be better because you tried lots of stuff," he said.