The Air-Speed Velocity of Unladen Gaming

When games are unexpectedly good

By Brad Wardell

For this column, I am going to have to plug what we’re doing but only as a mechanism of self-punishment.

How many times in the game industry have we seen it where a really cool game comes from an unlikely source? With the benefit of hindsight, I’ve regularly said “Oh, of course that’s going to be popular.” But it turns out that one can be surprised even when it’s right in front of his nose. You see, I’m in a lot of trouble. Big trouble. Not the organs-being-removed-while-strapped-to-a-table kind of trouble, but still big trouble.

Let me explain. Our company, Stardock makes operating system (OS) extension software, also known as desktop extensions. We also make games. But realistically, 90% of our revenue comes from the OS extensions and 10% from the games. It’s not that the games don’t make any money, it’s just that OS extensions make a lot more and at a far lower cost.

Consider the costs of some of these games. One recently released mega game took 3 years and $6 million to develop. It’ll be a big hit I predict because it is a great game with massive distribution. But it’s not like it’s going to turn around and make $100 million. And it’ll make most of its revenue in the next 180 days. Since there will be a console version of that game, it’ll probably make quite a bit of money (more than $6 million). But the risk…my goodness, the risk! In contrast, an OS extension can typically be done by a team of less than 6 people in about a year or 18 months tops. It can be sold to just about anybody and it can sell in high volumes.

So when it comes to assigning developer resources, we tend to give the majority of our attention to the things that pay the rent. It also provides me with the luxury of talking about games; since it’s not the only thing we make, I can be frank in my discussions about the industry without worrying about offending retailers, the media, or anyone else.

When we bring on new people to work on coding, rather than hiring people with years of experience, we usually hire people very young and then train them. We’d rather find that diamond in the rough. The first project that they’re usually assigned is something fairly trivial. We used to have them do a Pong type game just to get used to dealing with the whole checking in/out code, working on a team, debugging, etc. But for whatever reason, when brought on a new group of interns last summer, we gave them a throwaway project. It was a side-scroller, a la Super Mario Brothers.

The idea was to let them learn on this game. Then, once they mastered our various technologies, we would put them on a 'real' project. We even let them come up with their own story, title, etc. for the game. They came up with the title "LightWeight Ninja". And they came up with a plot that I thought was pretty weird – a genetically enhanced assassin who ends up going against his masters. If the game was a 'real' project, that plot might have been killed (there was no Dark Angel TV show yet). But hey, it wasn’t going to be a real game, so they could knock themselves out.

The game library they were to work with is the same one used for The Corporate Machine and the same one we’re using on Galactic Civilizations. Both are strategy games. LightWeight Ninja, on the other hand, could be described as a modern version of Commander Keen.

I need to talk about the game library for a second so that we’re on the same page. We have a team that creates a game library called Pear. Pear is designed to handle the low level sprites, GUI widgets, tile mapping, sound, music, and all the other stuff that games need. Pear is an incredible technology. None of our games thus far have really shown off what it can do.

But the interns, working on their practice game, ended up putting together quite a neat game with Pear. For instance, have you ever played a side-scroller on a PC? They just never seem quite as fun as they do on a console because you can see all those pixels. On a TV screen, the blurry picture acts as a natural anti-pixel mechanism. But the kids solved this by using Pear’s alpha blending features. In effect, even though the game is a 2D-cartoon style game, it effectively has full screen anti-aliasing. But since it’s on a per object basis, there’s no performance hit; you can run the game on a low end machine just fine with 30fps. So what you end up with is a fluid cartoon looking game without the pixels.