Current planning has a "star-topology": urban office core with rings of gradually less dense, gradually more residential development. This creates a natural transportation bottleneck at the center, and creates areas that are "alive" during only one part of the day, which makes it hard for local merchants to thrive (standard Jane Jacobs).
We can define a "balanced area" as having roughly balanced proportions of residential, office, retail, schools, hospitals, (light) industry, etc. The effect of an area being balanced is that the net flow of people across the areas boundary at any given time of day should be roughly zero - if you live there but work elsewhere, then you need to commute out, but someone else needs to commute in to fill the job in the area that you're not occupying. Same for other uses. "Area" could be anything from neighborhood, to part of city, to city, to region, to state - the analysis is roughly the same regardless of geographical zoom level.
Hypothesis: you can create a scalable city by constructing it around balanced areas.
Scalable = can expand indefinitely-ish. While not precisely true, it's probably a lot more true than for the current star topology.
So what about public transit? The reason we have the star is after all (in part) to make it possible to use public transit. Answer: increase density and switch from star to grid.
The star is hopelessly plagued by having everyone going in the same direction at the same time. This means that we're using roughly half (60%?) of our transportation capacity at any given time.
If we build our cities as dense, balanced cells (= 0.25-0.5-1 mile by 0.25-5-1 mile chunks, depending on density and other design criteria), with grid-oriented mass transit (buses, light rail, subways, whichever, all of the above), then we will naturally double our transit capacity (people going in both directions at all times), and shorten the distance people need to travel to get where they're going. The relevant metric isn't so much "miles per hour" as [density of people]*[linear distance per unit time].
Depending on the level of density you're willing to go to, different tiers of public transit will unlock: at lower densities, you can only really rationalize buses. Medium density = light rail, high density = subway.
Some additional points:
- Zero surface parking, but plenty of drop off / pick up spots.
- Organize the balanced cells around center squares, which are pedestrian only. Squares are focused around restaurants / retail / commerce; they're not parks, though they would likely have some trees. Make it the primary public space. Regulate building heights such that there's limited shading of the square.
- Smaller, pedestrian-friendly blocks around the center squares: stop signs for cars, no thoroughfares to other cells.
- Variable height zoning: stop making neighborhoods uniform. Allow vastly different heights within short distances of each other. This will create both views and natural variation in land and housing value = naturally more inclusive, and much more interesting.
- Turn the towers: don't build two 700' towers across the street from each other. Instead, space them out more, and rotate them so the tower doesn't face the street, to create longer corridors of views from the tower windows, and an "airiness" in the city.
- Between cells: focus more on surface traffic, limit pedestrian use. Yes, this will create "islands between traffic corridors". That's ok - zone light industry, gas/recharge stations, parking structures and office towers on these corridors and zero residential.
- If you're willing to up-zone and build as of right, you can start lower density (= buses), then as the city grows, you can up-zone and graduate to light rail, and then if you go truly dense, retrofit a subway. If it's in the city charter from the start you might be able to fight off the inevitable nimbyism.
- A pragmatic sister project to the study of cities would be the study of how to make public transit projects not cost so much. Taboo subject, and perhaps "boring", but super-important. If it cost half as much, would we have four times as much of it?
We can define a "balanced area" as having roughly balanced proportions of residential, office, retail, schools, hospitals, (light) industry, etc. The effect of an area being balanced is that the net flow of people across the areas boundary at any given time of day should be roughly zero - if you live there but work elsewhere, then you need to commute out, but someone else needs to commute in to fill the job in the area that you're not occupying. Same for other uses. "Area" could be anything from neighborhood, to part of city, to city, to region, to state - the analysis is roughly the same regardless of geographical zoom level.
Hypothesis: you can create a scalable city by constructing it around balanced areas.
Scalable = can expand indefinitely-ish. While not precisely true, it's probably a lot more true than for the current star topology.
So what about public transit? The reason we have the star is after all (in part) to make it possible to use public transit. Answer: increase density and switch from star to grid.
The star is hopelessly plagued by having everyone going in the same direction at the same time. This means that we're using roughly half (60%?) of our transportation capacity at any given time.
If we build our cities as dense, balanced cells (= 0.25-0.5-1 mile by 0.25-5-1 mile chunks, depending on density and other design criteria), with grid-oriented mass transit (buses, light rail, subways, whichever, all of the above), then we will naturally double our transit capacity (people going in both directions at all times), and shorten the distance people need to travel to get where they're going. The relevant metric isn't so much "miles per hour" as [density of people]*[linear distance per unit time].
Depending on the level of density you're willing to go to, different tiers of public transit will unlock: at lower densities, you can only really rationalize buses. Medium density = light rail, high density = subway.
Some additional points:
- Zero surface parking, but plenty of drop off / pick up spots.
- Organize the balanced cells around center squares, which are pedestrian only. Squares are focused around restaurants / retail / commerce; they're not parks, though they would likely have some trees. Make it the primary public space. Regulate building heights such that there's limited shading of the square.
- Smaller, pedestrian-friendly blocks around the center squares: stop signs for cars, no thoroughfares to other cells.
- Variable height zoning: stop making neighborhoods uniform. Allow vastly different heights within short distances of each other. This will create both views and natural variation in land and housing value = naturally more inclusive, and much more interesting.
- Turn the towers: don't build two 700' towers across the street from each other. Instead, space them out more, and rotate them so the tower doesn't face the street, to create longer corridors of views from the tower windows, and an "airiness" in the city.
- Between cells: focus more on surface traffic, limit pedestrian use. Yes, this will create "islands between traffic corridors". That's ok - zone light industry, gas/recharge stations, parking structures and office towers on these corridors and zero residential.
- If you're willing to up-zone and build as of right, you can start lower density (= buses), then as the city grows, you can up-zone and graduate to light rail, and then if you go truly dense, retrofit a subway. If it's in the city charter from the start you might be able to fight off the inevitable nimbyism.
- A pragmatic sister project to the study of cities would be the study of how to make public transit projects not cost so much. Taboo subject, and perhaps "boring", but super-important. If it cost half as much, would we have four times as much of it?