A Commuting Principle That Shaped Urban History

Overall I liked this article. It describes an interesting though simple core principle. I have three criticisms:<p>1. The article is scientific enough that using &quot;grew exponentially&quot; in the colloquial sense of just &quot;grew a lot&quot; is grating.<p>2. The article gives the impression that Boston is an American city on a grid where elevated trains preceded subways and dominated the transportation system. In reality the core of the city was not at all a grid, subways came first, and subways have been more prevalent than elevated trains.<p>3. &quot;the vaunted self-driving car, as imminent and yet illusory as nuclear fusion, will not transform the basic geometry of road capacity&quot; This ignores the potential drastic reduction of space allocated for parking, including space that is already part of roadways.

&gt; <i>The best option is to densify our cities. This is hard, too</i><p>Leveling off or lowering the human population through lower birth rates works better. I&#x27;m not saying it&#x27;s simple or even possible since so many people jump to conclusions about eugenics etc, but reducing overpopulation makes a lot of problems easier. If we don&#x27;t figure out how on our own, nature will do it for us so we might as well figure out how in a peaceful, fair way.

Except that people didn&#x27;t used to &quot;commute&quot; in the modern sense: many craftsmen worked from home, and large numbers were domestic servants.<p>There is probably some other principle at work, of which modern commuting is a special case.

&gt; the vaunted self-driving car, as imminent and yet illusory as nuclear fusion, will not transform the basic geometry of road capacity<p>I&#x27;m not so sure about this. I&#x27;ve generally been someone with a lower tolerance for a long commute than the average person. Usually I&#x27;ve lived within a 20 minutes walk from my job. My wife and I have had flatmates for 4 out of our 5 years or marriage. I decided 2½ months ago that it was worth moving farther out in order for my wife and I to get more privacy. A big reason for this was that a friend who lived in the area told me that he usually got a seat on the line into central London. Indeed, having figured out an ethical way to <i>always</i> get a seat, my commute is significantly improved by my newfound ability to get into a writing habit.<p>If driverless cars or driverless double-decker busses become widely available and many people have the ability to get one-shot commutes, then I suspect that more people will ride them.

This agrees with my experience. I grew up in the centre of Madrid and I would always estimate around 30 minutes to get to any other part of the city by public transport.

For a while, I have occasionally considered the idea of turning the shopping mall inside-out. Rather than one building in the center, with a vast expanse of parking lot all around the perimeter, put the parking structure (and auto-service businesses) in the center, and place the shopping, offices, and restaurants around the edges. Then use elevators and horizontal transport systems to move people around quickly without needing their cars. Access from the outside would be by pickup&#x2F;dropoff cars, public transport, and pedestrians, and building facades would be immediately adjacent to the street.<p>I don&#x27;t think you can go straight from an everyone-in-a-car culture to a walkable+public transit culture in one step. So first you get people to put their cars in places that are less in the way, by driving <i>past</i> their destination and parking <i>on the other side of it</i> from the main transit network connections. This allows things to be closer together for pedestrians and the vehicles that do not park.<p>Imagine the degenerate case, where a worker in an office park wants to visit a store in the mall next door. When the office is surrounded by parking, and the mall is surrounded by parking, a direct walk requires traversal of two large parking lots, and the worker has an incentive to drive, from a parking space near their entrance to the office building, to a parking space near the mall entrance nearest their store. Turning the shops and offices inside-out, in the best case, puts the office and the shop directly across the street from one another, and the worker can cross on a pedestrian overpass faster than they can reach their car. In the worst case, the worker rides two ring-avators and two elevators, once to get to an office-mall adjacency point, down to street level, walk across, back up, and then ringwise again to get to the shop. To drive, the worker would have to walk into parking to reach their car, drive out to the exit, drive in to the mall parking, and walk out again to the shop. That&#x27;s not entirely eliminated, and would likely occur when the offic and mall are further separated, but it does make walking directly slightly more attractive.

I liked the article, but I did wonder if the author understood what self-driving vehicles would allow for, providing that they had maybe 7-8 nines in reliability and safety (if that is even possible).<p>People have already mentioned the idea of &quot;sending the car back home to park&quot; and &quot;being able to do something else in the car while commuting&quot;. But there are other things the cars could do:<p>1. Trains of cars - convoys - drafting each other, reducing energy use, while also increasing road density (bumper-to-bumper traffic - but at highway speeds).<p>2. No more &quot;traffic waves&quot; on the freeway, nor &quot;rubber necking&quot; at accidents, both of which likely contribute to the congestion issue.<p>3. No-stop intersections - again at high-speeds - because cars could coordinate&#x2F;negotiate their speeds and lanes to allow for &quot;collision-free&quot; operation (unfortunately, there will still probably be collisions - and they wouldn&#x27;t be pretty).<p>All of these options - and possibly others I haven&#x27;t thought of - rely on things that self-driving vehicles don&#x27;t yet have - mainly car-to-car communication - so all cars know where other cars are (even if they aren&#x27;t within sight-lines of the sensors) and how fast they are moving. Basically a form of swarm robotics, with constraints on motion vectors.<p>Of course, #3 (and maybe the other two) might not come about unless the chance of an accident drops to infinitesimally small values - perhaps it may not even be possible. Better safety systems for the occupants would likely need to be developed as well (for that &quot;impossible&quot; chance it would be needed). Such cars might also need to be &quot;trained&quot; on &quot;how to crash to protect occupants&quot; and&#x2F;or how to coordinate a collision with another vehicle(s) in real-time as the collision unfolds, to minimize occupant danger - this again, would require a standardized mesh network and coding to be able to take advantage of all the sensor data that would flood in (and out - as sensors and actuators become disabled by the accident unfolding - maybe the other cars involved - or not involved - could share their information &quot;from the outside&quot; to allow the disabled car to take better actions or whatnot). It&#x27;s actually a fascinating problem when I think about it - extremely complex, lots of parameters and such; I wonder if anyone is working on it? If not - they should...

I&#x27;d love to build a large subdivision with the streets laid out in a fractal curve:<p><pre><code> _ _ _|_ |__ __| | | | +---------- |__|__| | | _ _ _ _ |___ ___| _|_ | _|_ | +-------------- _ _ | _ _ |___|___| _|_ _|_ __ __ __ __ |_ | _| |_ | _| |_____ _____| _ | _ | _ | _ |__|__| | |__|__| | +================ __ __ | __ __ |_ | _| | |_ | _| |_____|_____| _ | _ _ | _ |__|__| |__|__| </code></pre> It would likely suck to live there, but the traffic patterns would be cool...

&gt; We must make do mostly with building up and densifying the urban areas we already have. As transportation goes, so go our cities.<p>If we make transportation better, our cities can be less dense. So can&#x27;t we make transportation better?<p>What causes congestion? Car traffic. Why does car traffic cause congestion? It&#x27;s an inefficient use of roads. What are some alternatives?<p>1) Build more roads. Sure, except it&#x27;s expensive, time consuming, and you quickly run out of land, and would have to build them one on top of another.<p>2) Fit more people into existing roads. How? Buses, trolleys, trains, subways, vans. Build out the mass transit to the suburban enclaves.<p>But all this supposes an urban center is also the only answer. What do people need to congregate in one place for? Usually to work in one place. But why work in one place?<p>Due to division of labor, most of us don&#x27;t do things with multiple completely different people on multiple completely different subjects; we&#x27;re just not that collaborative. But you may need to transport <i>your work</i> to another group in a company. At a bank, all this is paper; in a car manufacturing plant, this is car parts. The former is done digitally now, and the latter could be done, again, by increased transportation efficiency.<p>All the other things we use are already decentralized outside the city; schools, churches, supermarkets, health care, water, power, recreation. It&#x27;s really just business, and the ability to work decentralized, that constricts where and how we live. Transportation will make it easier to decentralize. And we can start by giving up our cars.<p>The other aspect of work decentralization I&#x27;ve already seen at a rental car facility. A computerized kiosk displayed a remote worker on screen, who processed my driver&#x27;s license, credit card and rental information and directed me to my rental car. It was surprisingly trouble-free and pleasant, and only one of us needed to be there.