Samuel Arbesman is the scientist in residence at Lux Capital, a science and technology venture-capital firm. He is the author of Overcomplicated:Technology at the Limits of Comprehension.
What's lost when anything is possible.
I imagine that most people who read Atlantic Cities love SimCity (and will go wild when a new version is released in March). What about trying SimCity in the physical world? Even if you're not an urban planner or mayor, this is possible, at least sort of.
It can be done by combining SimCity with another geek love: LEGO.
There is a growing tradition of building tiny cities out of LEGO bricks. I first discovered this when reading the amazing book The Cult of LEGO. Within its fascinating pages is a discussion of the use of micro-scale. Much can be built at this size, from Yankee Stadium to tiny fire trucks, to, yes, cities.
For example, there is Shannonia, a growing tiny city from LEGO constructed by Shannon Young. Containing skyscrapers amidst water and mountains, it is a futuristic cityscape of beauty. And it can easily fit on a tabletop. Here’s another city at this kind of scale as well.
For those who wish to be a bit more collaborative, there is even a micro-city standard for building interlocking city blocks to make sprawling beautiful cities, known as Micropolis. For example, a standard brick’s height is equivalent to nine feet, and all sidewalks are one LEGO stud wide. This standard allows creators to build LEGO buildings and streets, snapping them together to create neighborhoods and towns.
I myself have even attempted such constructions (see above), using my limited collection of LEGO bricks. And I have discovered that such micro-cities are not easy. Ensuring a level of detail at such a tiny scale (and with relatively large pieces) is no easy task.
But these sorts of creations (successful or not) raise an interesting question: how do we make sure that these cities look city-like? When we construct a city in SimCity, it has a historical path to it. We don't just say "I want to build midtown Manhattan." You first have to build a small town, make sure it functions properly, and then slowly grow it, adding bit by bit. You are making choices based on the city as it is, with all of its problems and virtues, from financial difficulties and crime sprees, to available land. The resulting city often has an organic nature to it -- a bit haphazard with that "city feel" -- as opposed to looking like it was constructed whole cloth.
This is known as path dependence, where one’s current state is highly dependent on where one came from, and the decisions one has made along the way. Essentially, the state of the system is dependent on its path through the space of possible states. Choosing where to start on a map, for example, or in which direction to expand, can have important implications.
Path dependence is an important idea in chaos theory and complex systems. But with LEGO (or any other building toy), there are no such limitations. And therefore, something can be lost. If all is permitted, cities simply don't seem as city-like.
So how to overcome this? Whether in the world of LEGO, or even the virtual world, there are tricks to create systems that have the feel of cities. This involves techniques derived from the world of computer graphics, and it's known as "procedural city generation." If you want to make a complicated city map, but don’t want to draw it by hand, or define it explicitly, there are a variety of ways to procedurally—or, algorithmically—generate a map, or the structure of a building, or a skyline.
This paper discusses a number of ways to generate cities. These methods derive from the study of systems that have many parts interacting in a highly connected manner. And that shouldn’t be surprising, because that’s exactly what a city is. If you're feeling fancy, you can use something known as Lindenmayer systems, related to fractals, to construct building structures. And agent based models can help make road maps.
While the Micropolis standard allows for the construction of fractal-like buildings (which can provide structural variety), it creates city blocks that for many might seem too grid-like. If you do not desire a city modeled on Manhattan, and want something more similar to the road system chaos of Boston or London, I recommend using a combination of grids and radial roads. Using a program such as CityEngine, which can automatically generate a city's structure (while mimicking a certain amount of path dependence), then modeling one's LEGO construction on that, should work.
Follow these types of approaches and you don’t need the historicity of an actual creation. You can simply build it from the ground up, in silico or in lego.