Greg Costikyan has just posted a kick-ass blog entry on process intensity over at the Manifesto Games website.
Process intensity is a key concept in game design, defined about nineteen years ago by Chris Crawford. Greg correctly points out (as Chris did before him) that the industry has been going for the low-hanging fruit for the last two decades: data intensity. Graphics, sound, video. Things that, while not completely worthless, are not related to the inherent and unique qualities of electronic games as an entertainment medium.
To put it another way, we've been doing precious little actual game design; we've mostly been implementing new versions of existing game styles, but with prettier graphics.
But as I've argued, we've come to the point of diminishing returns along that road. In future, we have to do something else.
I will spare you the metaphor of ever higher low-hanging fruit and the evolutionary niche of the giraffe (it would be too close to Chris's koalas versus goats). But basically, the industry has painted itself into a corner, and the consequences are becoming clear. I am not a pessimist, so I won't say video games are dead or the industry is doomed, but I think it's obvious that the level of production values that are expected nowadays are hard to achieve, and that the costs are enormous - not just direct monetary costs, but the costs in terms of lost creativity, and lost customers (and developers).
Greg points out the crucial difference between procedural content and process intensity - basically, the former is a subset of the latter. I would go a step further and argue that procedural content, in and of itself, is a solution to production problem - how do you generate a ton of content without doing it by hand. (Arguably, it can also solve a technical problem - how do I get content to the graphics chip using limited internal bandwidth and memory - but meh.)
(And by the way: I could rephrase my recent post on 3D graphics to say that 3D graphics are more process intensive - they are procedural content - and thus give more freedom to the designer.)
In my opinion, the key thing that makes the procedural content in Spore cool is not that it's procedural, but that it's user-generated. The procedural part of it makes it possible to exchange user-generated content using relatively low amounts of data, but this is a technical / logistical issue. It is not inherent in the game's design, it just enables a key feature.
When people say that 'procedural content is the future of games', I think most of them are thinking of production issues, and you could easily spend another decade or two implementing all of the cool SIGGRAPH papers of recent history and procedurally generate clouds, landscapes, trees, people, you name it. And this is cool and potentially useful and I actually really dig that kind of stuff. But it's still eye candy.
Unless. You can use procedural content to enable shared user-generated content, a la Spore. But you can also link the procedural content to the actual game design, to the rules and simulations that constitute the core of the actual game. Perhaps trees, clouds and creatures change as a consequence of the player's actions. Perhaps they can be used by the player in order to achieve the game's objectives. As soon as the procedural content starts to become a part of the outcome that is affected by choice (Mark Barret's definition of true interactivity), new things become possible.
But you can go further. Procedural content remains a subset of process intensity, even if you link the knobs and levers of your clouds and trees to the algorithms of your actual game. I don't want to split too many semantic hairs, but, well, content. Doesn't that remind you an awful lot of... assets? Static data? Stuff that is... not interactive?
At their core, games are about interacting with dynamic systems, with rules, algorithms, simulations. Play is a form of learning and exploration (see Chris Crawford, Will Wright, Raph Koster). In my opinion, the most promising avenue of progress in interactive entertainment is in the processes at the core of gameplay. (The use of physics in games is a potential example of this, but too often physics are a part of presentation, rather than interactivity.) Deepening the processes at the core of our games will make them richer.
More and more games track a player's reputation with certain groups (think Fable or Frontier Development's The Outside), but single-number-per-group reputation systems are really simplifications of individual agents with memories and social abilities. If many games 'tell' a story, and stories are, fundamentally, about people, then surely interactive storytelling should be about simulating people? Not using 'true AI' or speech recognition, but with simpler techniques that make NPCs into believable characters that players can interact with - in ways that are meaningful to the gameplay itself.
This is where I believe increased process intensity can have the biggest pay-off.