How I would run a game company.

This was written a year ago, and the only notion that I have changed is starting with a pre-built game platform rather than writing the first one from scratch.  By starting with a game platform the company would save a decade of time over writing the first one ourselves.

I would start by buying a decent set of core engines that could do First Person, 3/4 perspective, and overhead view.  Then I would use those engines with a set of scripting and graphic tools to allow designers to script and clothe a game at a very high level.  Ideally this high level scripting and clothing system will work the same as the engines are ported across multiple platforms.

I'd require that most graphics be procedurally generated on the fly from tiny description files.  If need be these resulting textures can be cached on disk after they are generated.  This will ensure that the games levels can be transmitted as rapidly as possible over the Internet.  Perhaps while someone is playing one level the next level is downloading in the background and is generating the images so that the level is ready to play as soon as the current level is finished.  Additionally these procedural designs could automatically add more details to the texture map as the game is played on ever newer, more capable hardware.

The game play should just fall out of how the engine interacts with the rules that are set-up in the game script.  Very little should be specifically programmed for any single game.  My game designers can request cool new features for the next release of the engine, but for current development, unless there is a show stopper, they will have to work within the design constraints of the current engine.

As the current generation of games are being built, my core team will incrementally upgrade the core engines to allow more features that have been requested by the designers.  Nothing will ever be totally thrown out and rewritten from scratch. This is very important. I want to write better games, but I want the change sets to be small enough that I can come up with new games every year.

My company would never do vaporware.  I would write new games on new engines every year.  Not a game that nobody sees for 10 years because the technology is rewritten from scratch every couple of years in an endless loop of broken promises.

Another possibility is that old games could be played with the newer engines, extending their lives.

These engines and design tool chains could be leased to 3rd party game designers to create an ecosystem of similar games.  I would just require the end user to download my core engine, and then they could download and play tiny little chapters that tell stories with awesome game play.

The core engine would actually be an entire game platform that interfaced with a web site to allow games to be purchased on the platform and to handle updates of the core, the platform, and the game modules.