What Went Wrong

As much as some things sound like great ideas, they often aren’t. Here are the top contenders for our biggest mistakes.

1. Too much, too soon

The spec for this game was more than a little overboard. It originally targeted the PlayStation, Macintosh, and PC. It had networking, supported 3D hardware, used motion capture, and contained support for just about everything else you could want in a game. It was also on a 15-month development cycle. To top it off, Zombie was writing the entire engine from scratch. On the up side, there was a healthy budget.

We dropped the Macintosh version right away, and we terminated the PlayStation version at alpha (when it was still running at five to eight FPS). Networking support was postponed for an expansion pack a year into development. While setting our sights so high was clearly the reason we got so far, dropping this many versions and features along the way was worse than anyone expected. I blame publishers as much as developers in this kind of situation. Publishers go through more product cycles than developers do, and should have some past experience telling them what is realistic to achieve. We learned the hard way what’s possible to accomplish in 15 months; as a result, we completed the game in 20.

2. Getting the team

The last thing that I expected was that it would be hard to find good programmers in Seattle. I was hired at the start of SpecOps’ development and didn’t manage to bring the entire programming staff on board for nearly seven months. Needless to say, this caused substantial delays. We had similar problems hiring the art team. During the first months, the employees we hired burned out because they were trying to accomplish the duties of several people. I’m not sure what can be learned from this, but planning around such problems in the future will save some headaches.

3. Losing the art leads

About a month before E3 ‘97 rolled around, our art lead and a senior artist decided to leave the company. I had designed most of Viper’s capabilities with them, and losing them at that critical time really devastated the project. With a little luck and some amazing dedication, we found people to fill in for them. We showed up at E3 last year with some wonderful art. Losing key staff just happens sometimes, and there’s little that you can do about it but pick your chin up and wait for someone else to come along.

4. Losing the publisher

Right around E3 ‘97, we heard that our publisher, BMG Interactive, was planning to go out of business. It was not clear if our game would make it to the shelves, although BMG tried to assure us that everything was fine. After many months of things being up in the air, Ripcord Games came in and bought the title. In the interim, however, we lost morale, and there was some misdirection in our work effort. Ripcord came onto the scene so late that it had to really rush to get a marketing campaign going. Nobody wants something like this to happen, but it comes with the territory.

5. Networking

As mentioned earlier, at one point we mitigated being over schedule by dropping networking capabilities from the release. I don’t think anyone actually believed that this was a good idea, but it somehow it happened anyway. The fact is that as much as 3D acceleration is the future, so is networking. Luckily, the gaming community has been taking it soft on us, and we’re working hard to get out an expansion pack that has a variety of network play options.

That pretty much concludes the walkthough of the major components of the Viper engine and the problems we had while creating the title. While I could only really touch on the major issues we faced and the solutions we devised, hopefully it’s enough to make the efforts of other developers a little easier. If there’s a component of the Viper engine that you would like to see explained in further detail, please let me know. I encourage other developers to describe their projects in similar detail, so that we can learn from each other’s mistakes as well as successes.

Zombie's SpecOps development team.

SpecOps: Rangers Lead the Way

Zombie Virtual Studios
114 1/2 First Avenue South
Seattle, WA, 98104
(206) 623-9655

http://www.zombie.com

Core programming team: 5; total contributing programmers: 13.

Time in development: 15 month development cycle, extended to 20 during development.

Intended platform: Initially targeted for both the PlayStation and the PC, with support for 3D hardware acceleration. Playstation dropped at alpha.

Critical hardware: P166 64MB machines with 3Dfx cards and two monitors.

Critical software: Visual C++ 5.0 and Sony’s development tools.

Notable technology: Viper engine, which was created for this game