Hacker News Reader

Comment by brlewis | Hacker News Reader

Overall I liked this article. It describes an interesting though simple core principle. I have three criticisms:

1. The article is scientific enough that using "grew exponentially" in the colloquial sense of just "grew a lot" is grating.

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.

3. "the vaunted self-driving car, as imminent and yet illusory as nuclear fusion, will not transform the basic geometry of road capacity" This ignores the potential drastic reduction of space allocated for parking, including space that is already part of roadways.

View on HN

> This ignores the potential drastic reduction of space allocated for parking

How so? If you mean because of more car sharing, I don't think that's going to happen. Most people would still want their own car. If you mean because of being able to let the car park itself outside the city, or even drive home: That would drive up cost for the owner and reduce convenience (having to "order" it 30 minutes ahead of time). In addition, it would worsen congestion on streets due to cars moving more.

I can imagine NYC with 1/10th the private cars and 4x the shared/rental/taxi cars, and traffic flowing much better than today owing to the additional bandwidth freed up by the 70-80% reduction in street-side parking space need.

NYC residents just might accept the trade-off.

As an NYC resident, I would much prefer to just have 1/10th the private cars, twice the number of trains (or whatever is necessary), and a massive reduction in the number of cars sitting around unused, with the resulting extra space being converted to green space, pedestrian-only areas, and bike lanes. Self-driving cars and the infrastructure they necessitate would make NYC more hostile to people, not less.

As a non-NYC resident, I would love to see the same thing in my town.

I don't want to piss away $40,000 + another $80,000 on maintenance and insurance and parking fees and bridge tolls on a self-driving car that will stand idle in my garage, or my employer's parking space for 23 hours of the day. I want better rail. I want more buses. I want to be able to ride a bike, without putting my life in the hands of lunatics who in any sane world, should not have been licensed to drive.

Surely, we could deliver all that for the money we are currently wasting on automobiles?

I doubt you could deliver better rail for that money. Rail is fantastically expensive, and if you get the routes wrong, that's a forever mistake. Buses are another story, but even NYC, which has the density to support a fantastic bus system... has a crap bus system.

Owning a car in a metro area is ~$8,000/year. There's 1.3 cars/household, so we're looking at ~$10,000/year/household.

My metro area's per-household budget is ~$17,000/year. ~$4,000/year on that is spent on transportation, including public transit, roadway, and highway maintenance and construction.

Tripling the transportation budget, at the expense of personal automobiles would absolutely deliver many of those things.

NYC has a crap bus system, because they have to share roadways with automobiles, which are the cause of NYC gridlock. If buses had the right of way, they would work much better. And it doesn't take being Nostradamus to figure out at least a couple of good, new train routes - in any metro area.

I'm sure you could find me some examples of train routes that were built, and then turned out to be under-utilized... But I feel that they would be the exceptions that prove the rule. Every major metro area is seeing population growth, so it's not like people will suddenly move away from the train lines. Not to mention that every line of rail tracks has the throughput of a super-highway[1], at much smaller footprint and cost (If you take into account the full cost of vehicle ownership.)

[1] In Vancouver, any one of the three SkyTrain lines has a throughput of ~14 lanes of highway, during peak hour...

I wouldn't exactly say that rail is fantastically expensive. Acquiring dedicated right-of-way for any form of transportation (rail, bus, car) through a developed area is fantastically expensive. But metal rails and overhead wires cost about the same as 12 feet of asphalt. And in terms of capacity, one track of rail moves +10k people per hour, while a lane of highway with personal cars moves about 2k to 3k [1]. Bus Rapid Transit is a nice middle ground because it has a little less capacity than trains but can be installed in existing right-of-way - like a road with 2 lanes in each direction or a lane of parking. So to me, better bus networks are about optimizing existing infrastructure. However cities still need to engage in new infrastructure projects as they grow, and it's often hard to beat the cost per capacity of rail in a new project: A line in each direction get's you the capacity of a multi-lane highway in a fraction of the real estate.

And in terms routing railways - it's really not that big of a deal. If you create a fast and economical route between two places, market forces tend to move people along that route eventually. Now you might have decades of an unprofitable line before you build up enough density to really take advantage of all that capacity, but it will happen... assuming it maintains support from politicians.

Rail is not a short-term solution, but it's a great investment in the long-term health of many cities.

[1] http://ccag.ca.gov/wp-content/uploads/2014/07/cmp_2005_Appen...

Train infrastructure is simply not economical by comparison to buses. You could have a bus system as good as, say, Buenos Aires, if you were willing to let it be run by the private sector. Cars are not entirely replaceable by trains and buses. But if you replace N privately-owned, mostly unused cars with N/10 privately-owned _shared_/rental/taxi cars whose utilization is much more than 10x that of the average privately-owned non-shared/rental/taxi car in the city, then you win big even if you also improve buses and even trains.

I would be happy with more buses or trams or any other kind of mass transit option. As long as the space taken up by private cars is drastically reduced and the resulting space repurposed for actual public use. Public space is currently completely wasted by being disproportionately allocated for cars, many of them completely unused and taking up space for free.

Buses and trains may not completely serve all the needs which cars can, but almost all of the remaining transit needs can be easily taken care of by bicycles, which should also be a much higher priority in the city. Either way, it really is a no-brainer that we must de-prioritise cars.

> Most people would still want their own car.

I don't think that's a given. The car is typically someone's second largest or largest expense, yet has only a 5% utilization rate. It will become much cheaper to pay by the ride rather than own both because the owner can get a significantly larger utilization rate and because monopsony, or at least fleet buying will reduce purchase and maintenance costs (part uniformity, wear leveling etc).

There will be transitional effects of course, as people currently 30-60 yo will have the hardest time "letting go" but the economic value released will be significant.

Self-driving cars will not go outside of the city to park if it is more time and fuel-efficient to just circle within the city to avoid paying for parking.

Self-driving cars will get smart and avoid circling; they will park illegally as they please, so to speak, and then move when someone shows up to ticket or tow them away.

> This ignores the potential drastic reduction of space allocated for parking, including space that is already part of roadways.

Except that if a car isn’t in a car park then it’s on a street moving (circling, going to pick up a passenger, etc) and contributing to congestion (maybe more so for that reason as some other articles have speculated). So basic geometry of road capacity still applies.

If a car that would otherwise be parked for 8 hours in the city can instead drive 30 minutes back to the suburbs, park for 7 hours, then drive back in, we can trade 1 car-hour of freeway and 7 car-hours of suburban parking to regain 8 car-hours of city parking. And maybe we think that suburban parking is 'free' because it would have otherwise been sitting empty.

Obviously within the constraints of that model it's still debatable whether that's a good deal, between the emissions and the doubling of near-rush-hour freeway traffic.

And beyond that simple model, how cars performing pick-ups will impact congestion and whether people are willing to summon their cars 30 minutes in advance is also debatable.

If every car did two daily round trip commutes into the city center instead of one, that would double congestion. I just don't see this being workable at all.

I don't like the idea, but I've noticed that in my city, if you're heading into the city centre you're absolutely out of luck as far as traffic is concerned, but if you're heading the other way, it's extremely light. So you wouldn't so much be "doubling" congestion as just applying it to both in and out flow equally. I don't know if that would actually cause more difficulties or not.

Not everybody is going to the exact same spot downtown, and there's also some actual reverse commuting, so there will be some overlap between the to-work and from-work trips, especially where there's the most congestion (I e. Downtown).

Absolutely, I just don't don't know how much that will increase congestion. I don't think it would actually be a 2x worse experience due.

It's really not light going the other way, not in big cities. Try looking at the traffic patterns in NYC or LA. It's just bad traffic all day long.

Except that on the departure trip after dropping off and the arrival trip before picking up does not have to take congested routes and instead take paths that could reduce congestion. It's not like the car has somewhere to be during non-peak hours.

Unless reducing parking incentivises people to not have a car in the first place.

If that’s the case you can reduce parking and incentivize reduced car use regardless of automation. It’s just an urban planning matter.

It doesn’t follow that automation reduces the need for parking which then causes people to not own a car because there is reduced parking.

Hmm... but if someone only actually need a car occasionally then there's a saving. Automation could enable car-sharing.

I think "growing exponentially" is technically correct. Anything that grows at a rate proportional to its size is growing exponentially.

The transportation innovations allowed cities to keep doing that, i.e. growing at X% per year instead of hitting some limit where population growth would force out-migration.

OK yes there's a reasonable use of the term early in the article that I overlooked. It was the "streetcars and bicycles" section where it grated on me.

This may be one of the rare usages of 'exponential' in the popular press that I can allow. The exponential aspect of urban growth may be apt -- the premise that the number of people who can find a reason to relocate to a city will vary in direct proportion to the number of people who are already there might hold with acceptable accuracy over some range, and departures may be negligible during some periods of rapid growth.

Reply on news.ycombinator.com