Darth Hater

Dalton showed some case studies from the development of The Old Republic to explain his problem-solving methodology. The game is being developed using the Hero Engine, an MMO engine which allows for real-time simultaneous editing by all disciplines -- great for visualizing changes, but put under strain by the size of the team working on TOR. Dalton was careful to note that his team stopped taking code drops from the engine providers some time ago and that the engine's capabilities may be profoundly different today.

At BioWare, writing is the foundation for their games. Problem: the communal editing server didn't work at first, and the writers had to be peeled off and given stand-alone tools -- creating unexpected complexities.

The server ground to a halt again later due to a problem with the sound design section. Instead of a complicated solution, "We asked what it would take to limit the size of the sound banks," says Dalton. The response? "'50 megs? Okay'" It was a painless solution -- or as Dalton put it, "They don't all have to be mountain-movers."

The game also implements Morpheme for animation. Says Dalton, "The problem here is that they're changing the animations live, and if they break things, it triggers a lot of fallout. So you'll have scripts that are now starting to fail, and you'll have NPCs that won't appear in the world. This was a nasty one because people weren't even aware why it was breaking."

The engineers briefly considered making the animators work on a different shift when others weren't using the editor. The solution reached, however, was simply changing process: the animators have to email the entire team when they're about to make changes. "It was causing so much havoc across the team but with these steps it's a non-event now," says Dalton.

At about 80 people editing simultaneously, the engine's editing environment started to show serious strain. "None of [the problems] are huge but they're all sand in the gears, slowing everybody down," says Dalton.

This change was major: the team had to branch the engine for different disciplines. "This was a massive undertaking, a huge change. What it enabled the team to do was spread out a little bit. We got pushed into a do-or-die situation with this one."

When asked if the communal editing environment was worth the complexity it brought to the development process, Dalton found it hard to answer. "I think that there's no single answer for that. I think it's not worth it at the team size we have now. I think it was absolutely worth it when we licensed the product and started work, because that's what you wanted in the early days."