Will gave a talk about designing user interface to simulation games back some time around '96, to Terry Winnograd's user interface design class at Stanford (at the time I worked with Terry at Interval Research), and I sat in on the class and took these notes:

http://www.art.net/~hopkins/Don/simcity/WillWright.html

He demonstrated an early version of The Sims, at the time called "Dollhouse", which he'd shown me an even earlier version of about a year before. At the time I was skeptical that he could get the AI to work, and I suggested he punt on the AI and just make an online multi player game. In retrospect, I sure was wrong!

In the talk, he discussed why previous Maxis games were successful, and gave a demo of "Dollhouse," describing what he thought would be interesting about that game, which eventually became The Sims. The key to the AI was putting the intelligence in the objects, instead of in the characters, and making it possible to plug in new objects with their own content and programming, to expand the game on an open-ended way.

The reason Will explained that SimCity was successful was that people already know a lot about the way cities work, so it's engaging, and it doesn't have to simulate details as much as just imply them, and let your imagination do the heavy lifting and colorful illustration. Computer games are much better at implicating than simulating, because any simulation is necessarily a vastly simplified caricature of reality, while your imagination is unlimited and has its own built-in "ai" and common sense knowledge base to draw on.

There are two important models involved in a simulation game: the sparse digital model in the computer, which stays the same, and the rich organic model in your brain, which grows and changes as you play and explore the behavior and limitations of the game. As you play the game, the computer is downloading the details of an organic model into your brain, which elaborates them based on what you already know, and links them into your global understanding.

It can be educational in that it makes you think about the issues you're dealing with, but you learn more by thinking about what you already know (and having your curiosity stimulated and being inspired to learn more), than it teaches you about what its simulation actually knows about. Instead of taking the simulation at face value as an accurate representation of reality, you explore and find the edges and limitations of the simulation, and compare them to reality. As you find the limitations of the game's model (in order to figure out how to take advantage of its limitations and cheat), you scoff at the game because you understand the world is much more complicated and nuanced than a computer game. But that's a good exercise to think about, and it stimulates you to discuss it with other people and learn more through other channels!

That is one aspect of what Seymour Papert and Alan Kay refer to as "Constructionist education": http://en.wikipedia.org/wiki/Constructionism_(learning_theor...

Anyway, SimCity was successful because it was about a domain everybody knows a lot about. And The Sims was even more successful since it was even "closer to home", so to speak.

But most people don't know much about plate tectonics, ant colonies, evolution or galactic conquest, so while SimEarth, SimAnt and Spore might teach you something about those topics and stimulate your curiosity, they weren't able to engage people and play off of their vast existing knowledge as much as SimCity and The Sims did.

But independent of how engaging or commercially successful it was, I think Spore was a very interesting and successful experiment in several important aspects of game design, which had to be done.

When Will described the early concept to me, of an online game with multiple levels that moved at different time scales, the first obvious problem I saw was that it would be impossible to coordinate the independent timelines in a massively multiplayer online game, since at each level, time moved at a different speed, so different players would be traveling through time at different rates, and it would be impossible for them to interact with each other in real time.

But Spore solved that problem by being an "Massively Single Player Online Game", where players shared content asynchronously, but didn't directly interact together synchronously.

I think the idea of sharing user created content asynchronously online is a great one, and Spore performed it very successfully.

The other important concept that dovetails with that was tackling in-game content creation tools. And I think those were a wonderful success, and paved the way for other games to do similar things.

With games like The Sims, it has some easy-to-use in-game tools for building architecture, but you have to go outside of the game and use tools like Photoshop to make skins, 3D Studio Max to make meshes, Character Studio and Biped to make animations, etc.

We made some simplified tools for The Sims (like Transmogrifier) that let you make objects by exporting and importing bitmaps and editing them with 2D tools like Photoshop instead of requiring 3D tools like Max, which opened up content creation to a lot of people, but making 3D objects still remained a very tough problem, and we never released the tools we developed internally to do that, since they were very hard to use and had a lot of environmental dependencies (like requiring expensive commercial products).

Building 3D editing tools into the game the way Spore did was very ambitious, and I think it worked extremely well. The user interface was very easy and fun to use in and of itself, and it gave players a huge amount of freedom to make everything from walking penises to hopping penises to flying penises to crawling penises, and even penis huts, penis houses, penis towers, penis planes, penis trains, and penis mobiles.

http://www.youtube.com/watch?v=LFOVYx90Ni8 http://www.youtube.com/watch?v=Cv-NjbyhXFo http://www.youtube.com/watch?v=puv6pwG_AZo

Spore had some great and successfully executed ideas about building and sharing user created content in it, but since it was an amalgam of several different games, and each sub-game was kind of like a tribute to an existing classic game, those games in themselves were not any better than the existing games, and they were not very well integrated.

Here are some notes I took from Will's talk at GDC, which I had him review for accuracy at the time:

http://www.donhopkins.com/drupal/node/35

You can compare it to what Spore actually delivered, and see the differences. There was a big gap between the ideal design as he articulated at that time, and what they were able to finally deliver. Of course that was inevitable, when working for a big company like EA, even with the freedom they gave him.

There was an overall high concept that the stack of sub-games was a ladder you would climb to get to the higher level storytelling based game, and that you would then swoop down into the sub-games to perform scenes of the story. (The "T shaped game".) But the sub-games were never consequentially integrated enough for that to work. So I don't think the storytelling game was ever fully realized.

That gap between the original design and the final product was of course because of the harsh constraints of reality and the necessity to ship something. They had to simplify it a lot, of course. And different people had ownership of each level, and the design decisions that one person would make at a lower level would spill over to the next levels as constraints and limitations they had to work within.

I think the essential problem of how can each level consequentially effect the other levels in a way you can make a high level storytelling game around is a very hard one.

The high level storytelling missions don't really require you to go back to the previous levels you climbed to get up to the galactic conquest level, and I have a hard time imagining how they could even do that. ("Go down to the protozoa level and conquer the evil amoeba who is about to sabotage the delicate diplomatic negotiations, by making the president throw up in the prime minister's lap!")