I'll usually only have the big picture of where I want to go and as the schedule dictates I'll flesh out parts of the game on the fly. Lock's Quest was a good example of iteration. I knew the combat system needed something unique, I had rough ideas of some stylus based slashing or attacking. When I actually got to play it, it was boring, it wasn't combat, it was a movie. You just stood and watched Lock automatically attack an enemy. It gave me the idea to add micro-touch games similar to Elite Beat Agents to enhance the gameplay which became a core feature of the game and turned out very well. I would have never thought of this system in a 300 page doc beforehand if I had written it. I had to be in the moment of the system to really come up with something unique.

Actually Scribblenauts had more like twenty-five to thirty people, which is kind of average for our DS games, but compared to AAA standards is still pretty light on DS. It's not very challenging for us to handle iterative design. For some strange reason we run our company a lot closer in style to a Japanese developer as opposed to an American one. Most American companies are very design by committee, too many cooks in the kitchen with no head chef, which can cause a lot of problems. Japanese companies are very visionary-focused. A leader or leaders dictate the flow, while people under them focus on their own systems under the guidance of the leaders, it's a trickle down system. This isn't micromanagement, it's just the people with the vision for the game drive it forward and hand out tasks to the others. Those people under the leaders then own the systems they're working on within the constraints of the vision. The leaders then check to make sure the tasks are following the main vision. Collaboration comes from iterating on the systems they're focused on, not the game as a whole, that's the job of the leads and directors. It's worked out very well for us. People like having a solid direction and knowing someone has that big picture. It's way better than going aimlessly for months because no one wants to commit to an idea.

How different is today's Scribblenauts from the original idea?

Pretty close. The original idea when Scribblenauts actually came together as a concept was write any object you can think of to solve puzzles. We've stayed true to that throughout. We wanted a sandbox mode, the title screen idea came later. It's kind of our little thing to have unique interactive title screens, all our games do it. The idea of breaking up the levels between casual and gamer also came a little later, but still early. We added wifi and level editing and sharing real late in the game. But as far as the core concept was concerned we saw it right off the bat and designed the game based on that.

Scribblenauts has a kiddy feel and some of the UI decisions were clearly made with inexperienced gamers in mind but it still draws an appreciative crowd of adults at parties. What audience was Scribblenauts designed for and what is your philosophy on designing for an audience?

The audience was everyone. Which is the hardest audience to make something for. It's very stressful to come up with concepts like that. It's so much easier to design for a genre or niche audience, even if that audience is huge like online FPS lovers. But doing a game on a system where half the demographic is kids and females and the other half is hardcore gamers makes design choices very difficult to please everyone. Ultimately you can't, but you try your best. The reasoning behind hitting this audience was motivated out of sales. We want to grow our company, we do one to two titles a year, we have no choice but to make hits. Flops mean no growth for us, and would affect our ability to remain independent. Even though it's bottom line driven, the great thing is we do very original, innovative concepts because we want to bring innovation back into the commercial industry, instead of it just thriving in the indie scene. And that goal and drive has succeeded for us consistently.

Did you play-test the game design (rather than just the UI) while Scribblenauts was in development? Did anything particularly interesting come out of the testing?

Play testing was very difficult because the systems weren't up and running until about 80% of the project was done, by then we had just a few months to polish the game. We were just learning how to use all the fun tools we've created and had little time to deeply iterate on the concept itself. This is the bane of all devs who are strapped to a publisher, which as far as publishers go WB is pretty darn good, they respected us a lot, were hands off, not trying to make their vision. They wanted our vision and were collaborative, not dictating. But still they have to hit the shelf at a certain time, so we just didn't get the time we wanted to truly play-test as deeply as we wanted which caused issues in the design. It's really hard to polish crazy unique ideas in limited timeframes. Usually that level of polish comes from a few iterations and sequels. Drawn to Life 1 (DS) vs Drawn to Life 2 (DS) is night and day for us in terms of polish.

I sat down with the Scribblenauts level editor to make some action levels. But when your player has access to everything on Earth it's pretty hard to make a challenging level. Can you talk about the difficulty of making action levels vs. puzzle levels and why the two separate classes exist?

Making levels for a game where people can write tens of thousands of words that could break your level instantly or come up with solutions you've never thought of is very difficult. Our designers had a very hard time thinking outside of the box to come up with 220 unique puzzles. With Action levels it's hard not to fall into old ruts of platform gaming or puzzle switches and gates type gameplay when as a designer your brain is trained that way. In puzzle levels the challenges where making logic that everyone could get, different regions, backgrounds and upbringings meant not everyone was familiar with all the nuances of certain cultures in a game that was going to ship around the world. So we had to design in very broad strokes which made narrowing down concepts difficult.

We separated the game into two types of levels because we wanted something that was logically challenging (Puzzle) for non-gamers and gameplay challenging (Action) for the gamers. You can play only puzzle levels or only action levels and still beat the game. The title screen, which is a complete toy sandbox for non-gamers, rounded out the experience.

I am fascinated by the decision to allow synonyms in the "three consecutive times" challenge. By allowing synonyms it seems you are telling the player "Yes, you can solve it the same way three times. But are you really going to? Where is the fun in that?". Which seems like it successfully blurs the line between game-based goal seeking and toy-based free play. Can you talk about this a bit?

That was a big discussion for awhile, allowing synonyms, do you allow people to cheat themselves or force creativity? Ultimately we came to the conclusion that if this game was targeting the widest audience possible, then we needed to allow people who can't think creatively to make something unique and not feel stuck. Again Scribblenauts really relies on people challenging themselves. I see on message boards people imposing their own rules to beat levels, only using levels that start with the letter A, or trying to beat the entire game with never using the same word twice. That's the textbook definition of emergent gameplay and I love to see it.