Why don't game developers do what I want?

The fervor behind the emergence of the first-person shooter ("FPS") game genre, began by Wolfenstein 3D and cemented by its successor, Doom, was in part because of the implicit promise: soon, I will be living in a virtual world. Heck, I'm already pretty much there, looking out through the eyes of a person as they have complete freedom of movement in a 3D world. Soon, I'll be in the cyberspace of Snow Crash. W00T.

That was 1993. Yeah, fourteen years ago. What happened? Why has the most significant interaction innovation in the genre been "jumping"? Why are "cyber-cafes" depressingly non-virtual places, where my consciousness manifestly does not exist in a luminiferous ether of an endless datastream? What went wrong?

There are a few of reasons why our "virtual worlds" evolve slowly: there are simple manpower issues, there are technology issues, and there are interface issues. In this post, I'll deal with the first reason.

Mostly, a world costs too much to make. A little over a decade ago, a game character in a AAA title could be painted by a single artist in a week or two. Now, it takes a team of artists months to model, texture, rig, animate, and script a single character. Back in the day, Non-Player Characters could be single fullscreen images of a person who talks to you via box of text, and sells you groceries or swords or whatever. Now, NPCs take nearly as much effort as main characters, and overall may take more, because you need a lot of variety in the NPCs that fill your virtual world. It takes a lot of time.

The same goes for the world: once upon a time, a simple backdrop painting of a town, or a tileset of walls and floors, or a collection of sprites, were all it took to make "the world". Now, artists find themselves modeling and texturing buildings, doors, doorknobs, door jambs, door mats, ten different door mats because you need variety, etc. The endless level of detail that we are accountable for is ridiculous.

The tools we have at our disposal are largely the same: a paint program, a modeling program, but we're painting orders of magnitude more pixels, modeling orders of magnitude more polygons. Whatever gains in sophistication our tools have made in helping an artist's workflow are more than offset by the much larger quantity of data that must be generated, and the greater accountability for detail that the refinement of graphics imposes.

What does this mean for players who see in their minds the game that so obviously needs to be made? It means to get the game they envision with current tools and technology, a company needs hundreds of artists and years of development, and one hopes the population of the Earth buys the game because that's what it will take to turn a profit. And this is just the Art part: similar issues exist for game systems and engine development.

Next time: Technology, Our Recalcitrant Friend.



IEEE Spectrum’s gaming blog was retired in 2010, but it is preserved here for archival reference.