In this reprint from the August 2002 issue of Game Developer magazine, Spider-Man developer Jamie Fristrom writes about what went right and what went wrong with the game's development process. Fristrom's current project, Energy Hook, strongly builds upon some of the design principles first explored here and in Spider-Man 2.
Treyarch was finishing up Max Steel and Tony Hawk's Pro-Skater 2 for Dreamcast when we agreed to do a game that would tie in with the Spider-Man movie, and release it simultaneously on all three of the next-generation consoles: Playstation 2 (PS2), Xbox, and Gamecube. We formed a new team out of parts of the others to begin work on the proof-of-concept and design. This team of four programmers, four designers, four artists, and a producer wasn't starting totally from scratch; we had Activision's previous Spider-Man game for the Playstation (PSX) to look at.
We wanted to improve some of the things about the game, such as giving the web-swinging more freedom, and play to that game's strengths, such as the hostage modes and variety of ways in which you could be Spider-Man. We also wanted to add a whole new type of gameplay: aerial combat, the ability to take on flying villains as you swing around Manhattan. Sony/Columbia gave us a ton of concept art and stills from the movie on which to base our work.
The PS2 was our lead SKU, because it has the largest installed base and was the easiest to get development kits for. We figured if we could make our game run on the PS2, we could make it run on anything. We began work on the Xbox a few months after the PS2 and the Gamecube a few months after that.
"All you need is good people," says the submarine commander in Das Boot, after the chief engineer repairs the boat and saves their lives. It's true, with good people you don't need to enforce process because it happens automatically. Everyone makes sure they do a good job: they volunteer for code reviews, they write their own unit tests, they find better ways to do things instead of the ways passed down from on high.
Our team was made up of a number of talented individuals, who each made their own unique, lifesaving contributions. Even the interns were amazing.
Because I'm not in HR, I don't know how we got these people, but I think part of the reason is that the people in charge of interviewing know their stuff, whether it's art or code. Part of it is the referral bonus we give employees for recommending new hires; part of it is that Treyarch is an environment not too many people are willing to leave. (We've seen what's out there, and it's worse.) Part of it may be the free soda, and free dinners during crunch time, and part of it is our policy of not hiring just anyone to fill a position. (We do hire out of desperation occasionally, but rarely. And when we discover we've hired someone who is just average, we let that person go.)
This team was not only talented but also motivated. Even though adding unplanned features was discouraged, many of us stayed late, on our own time, to get stuff in we thought would make a real improvement to the game. This is where the playing-as-Green-Goblin mode came from, and the secret bowling level, the rats in the sewer level, and a lot of special effects.
Treyarch was developing a few titles for the next-generation platforms, and it was obvious that having one crossplatform library to do the rendering, which all the teams shared, would be a big advantage. We formed a new team, Next-Generation Libraries (NGL for short). The Spider-Man graphics/PS2 programmer split off from our group to lead the team, providing the architecture and API to which the various platform graphics libraries would be written, and he developed the initial PS2 graphics library.
NGL had disadvantages as well as advantages: the advantage of more efficient use of coder resources was balanced with the worry that we might not be able to rely on a separate team. The NGL team wasn't always kept in the loop about when our deliverables were due and sometimes weren't available on a weekend or night when they were needed. Finger-pointing would occur: is it an NGL bug or a client-side bug? And sometimes finger-pointing didn't happen when it needed to: a client-side bug would linger on the plate of an NGL programmer who didn't know how to fix it.
At times, the NGL programmers became de facto Spider-Man programmers. They were building the Spider-Man code base on their machines to make sure their changes worked with it, to optimize for the game's worst cases, and to make sure the Xbox and Gamecube matched the PS2 close enough. In the end, NGL worked out great, finishing features and fixing bugs in time to ship.