Is parking too cheap?

Yeah highways are easy enough. Most drivers are on mental autopilot anyway, and a computer could do that just as well, plus react better at braking.

Except when a plastic bag flies by....

Ahem.

Again I point to my hybrid/electric car example. The basics are there. The product launched 15 years ago. But its a miniscule portion of the market because price and range havent been solved yet. And those are pretty damn serious problems, even though tons of money and thousands of very smart people are working on them every day.

And as great as Tesla is, 200 mile range wont cut it for most people, and even the $30,000 they keep promising is still way out of reach for most people (and keeps getting delayed...)

Even if we get a 500 mile electric car at $20,000 in 10 years, it would have been 25 years of real commercial experimentation (never mind another 2 decades of lab work) to even get the car to become a majority of new sales, and another 10 years to become a majority of the fleet.

These things take time.

Why not start with giving the MBTA self-driving trains?


This analogy doesn't hold since hybrid/electric cars have a hardware issue, while self-driving cars have a software issue.

Hybrid/electric cars are running against the laws of physics. There is only so much energy density you can pack into a battery before it spontaneously combusts. The marginal return on battery life decreases as energy density goes up (barring any sort of game-changing technological advancement).

There are some issues with self-driving cars' hardware, such as sensing stop lights when the sun is directly behind the light. However the main obstacle to overcome is how the car should react when something complex/unknown happens.
 
This analogy doesn't hold since hybrid/electric cars have a hardware issue, while self-driving cars have a software issue.

Hybrid/electric cars are running against the laws of physics. There is only so much energy density you can pack into a battery before it spontaneously combusts. The marginal return on battery life decreases as energy density goes up (barring any sort of game-changing technological advancement).

There are some issues with self-driving cars' hardware, such as sensing stop lights when the sun is directly behind the light. However the main obstacle to overcome is how the car should react when something complex/unknown happens.

Thats a good point about the hardware and software.

However, self driving cars are relying on a lot of technology too. Radar, 6 cameras, infrared, and a whole bunch of sensors. All that technology exists, but it will add costs and maintenance, which of course will delay the push of self driving cars into the affordable mainstream.

And yes, the question is, can we program self driving cars to deal with the unknown? Is that really something we can do in ten years? And can it be done in a zero error way? My internet browses still freezes every so often, which is annoying, but a car having a freezing issue as it approaches a busy intersection can be catastrophic, for example.

This country is still terrified of self driving trains, I just cant see self driving cars on the roads, again outside of highways, any time soon.

The danger is that there are already people rallying against transit proposals insisting theyre a waste of money because self driving cars will solve all the mobility issues in 10 years.
 
Anyone who rallies against transit 'cause of self-driving cars' is full of shit. You know that they're just anti-transit snobs looking for another flag to fly.

Self-driving cars might solve some problems. But they don't solve the most fundamental problem: geometry. Only so many vehicles can fit into a city designed for human beings.

To pick an image from today:
rutherfordtraf.jpg


Now imagine the same scene in some decades with self-driving cars:
rutherfordtraf.jpg


Different era, same problem.

(random note: I used to work at the same lab as the Chris Urmson quoted in the article you pasted, before he got his fancy new job. Not that it means anything, we didn't work together. Just coincidence.)
 
I'm getting way ahead of myself here, but in theory we could get rid of stop lights and traffic jams entirely once self-driving is sophisticated and ubiquitous.

Cars could communicate to each other with ad hoc wifi networks during a traffic jam to accelerate simultaneously. They could also communicate with each other about their intentions to turn/go straight at an upcoming intersection such that neither car has to come to a complete stop. I'm still not sure how we would work pedestrians/cyclists into that equation though.
 
As someone who rides a bicycle year-round in mixed traffic, I really hope they don't let self-driving cars loose in public any time in the near future. Too many crashes are already classed as "accidents" which would be completely avoidable if people just paid attention to driving while they are in their cars. I concur that self-driving trains seem like a no-brainer in comparison -- the things are on a constrained right-of-way, after all.

And back on topic -- yes, car parking is too cheap.
 
^ good to see this out there. Seems to me a per-neighborhood auction is the way to settle this, with prices rising (or re-set annually) to a market clearing price (where the number of permits issued equals the number of spots. Rich areas would pay more, easing the "regressive" aspects.

Unit owners get a tax rebate equal to half a car, or a whole Hubway membership, and the rest should go to amenities like heated bus shelters (winter) and bike share (non-winter)
 
I guess I have to disagree with Professor Shoup on just that one point there. I believe that there are plenty of American cities, not just Boston, that try to repeal the law of supply and demand with regard to parking. They always fail, of course.
 
The difference between smartphones and self-driving cars is that smartphones are engineering hard (which takes time and money to solve, but we understand everything necessary to do it) and self-driving cars are hard hard (which takes an unbounded amount of time to solve yet unanticipated problems because pattern recognition and dealing with arbitrary input is hard), not to mention politics hard.

I imagine there are security researches eager to try to "hack" self-driving cars into doing unanticipated things.
 
We already get 90% of the benefits of self-driving cars from taxis. They roam the streets serving multiple users and don't occupy parking spaces in high demand areas. Uber has proven (to me at least) that the problem with Boston's taxi system is too low supply, too high price, and the issues caused by municipal boundaries. Uber and Lyft actually solve these 3 shortcomings and anecdotally, taxi/Uber uses among myself and my friends is up about 1000% since Uber came to town.

What the taxi doesn't provide is a car in your driveway that you can take away from the dense urban center. For that you have Zipcar or traditional car rental. Nothing new there, other than changing the way people think about using rentals.

There is little we stand to gain from the whiz-bang self-driving car and I agree with the other engineers that have explained why we aren't going to get them. We will need to see AI functioning at a very high level on much simpler problems before we have self-driving cars.

What we should focus on is getting public policy regarding taxis changed and some public parking resources given over to shared vehicles instead of personal vehicles.
 
Google says its cars already drive better than humans, and are better at seeing cyclists and pedestrians because they can detect/track small & low-speed changes in the positions of peripheral objects better than humans can (in what is "out of nowhere" to us is a 360-deg model to them). Between LIDAR and Infrared cars are already better at seeing heedless fashionitas who insist on cloaking themselves in black at night and in rain. Google's cars are also more patient with weaving idiots avoiding risks that human drivers are liable out of anger-stupidity:

anigif_enhanced-31317-1398696428-23.gif

Note that the "omniscient" black-&-white view above is actually constructed from LIDAR, and anything not fixed gets colored red as a "mover" and the green "fence" shows extra effort on immediate/close movements. How much better would you drive if you could direct your car from the third person omniscient POV?

The other thing techno pessimists are missing here is that individual car owners, and city taxi companies don't drive "everywhere" and under "all conditions". We're finite people with finite time and finite interests and so we drive on a very finite number of streets. Just like Google knows whether PVD means Peripheral Vascular Disease or Plasma Vapor Deposition or Providence Airport depending on whether your other searches indicate that you're a diabetic, a diamond freak, or a frequent flier.

We're never going to have a car that anticipates all combinations of all road conditions (just like the Traveling Salesman Problem), but just like the TSP--WE NEVER NEED TO. Between clever approximations and just showing our cars "our" trips, we're basically there.

My car will learn, by me showing it, how to drive my VERY limited set of roads over and over and over. Which will totally nail our daily commute and 80% of our errand trips (which is why we're talking about it in the parking thread because parking is all about home, work, and errands). It'll also know everyone else's "popular" trips, but be ignorant of all unpopular trips, like a rural driveway, until some guy in a fully-instrumented AI car drives it for the first time.

No, out of the box, the very first self-driving car isn't going to know that the Jamaicaway trees impinge on the right, but it'll know that after just one guy shows his 1 car that.

This greatly reduces the problem of having AI that can learn to make city trips. Even without info-sharing, you'd just drive it home-to-work for a week by every way you know, and it'll learn it (as Waze does now at a grosser level)

We already get 90% of the benefits of self-driving cars from taxis. ...What we should focus on is getting public policy regarding taxis changed and some public parking resources given over to shared vehicles instead of personal vehicles.
Well, yes, for taxi use and use of taxi-like things to become more widespread, we need two things to happen:

1) Taxi-like modes need to get cheaper and more convenient.
- Self driving helps on labor costs
- Prepositioning them in parking spots helps reduce terminal time
- Ubiqity means they arrive faster (if not parked nearby to begin with)

2) The cost and inconvenience of competing modes (personal car) has to go up, by:
- Unbundling parking from housing, so parking costs can rise without driving up housing costs
- Raising the cost of parking, both off-street (by limiting supply) and on-street (by letting prices rise to supply-demand equilibrium)
 
Last edited:
From NBBJ Boston:


WHY I STILL OWN A CAR
Free Parking Creates a Perverse Incentive for Car Ownership

February 4, 2015
by Alan Mountjoy
Planner / Architect, NBBJ

...

So why do I still own a car? Did I mention that Brookline has an exclusionary practice of banning overnight parking on its streets? Thus the only real option in Brookline is either to park over the city line in Boston, to rent or buy a space, or to be so lucky as to have a deeded space on your property.

Which is the only reason I have a car. Because 15 years ago I bought a condominium with a deeded off-street parking space. It’s that simple.

If we want our cities to be more walkable, maybe it’s time we made them less park-able. When new developments provide parking, as current regulations require them to, they just encourage more people like me to own cars they don’t really need. Much discussion has been going on in Boston about developers building housing with little or no parking: existing residents fear streets clogged with cars parked (warehoused) on city property. Yet when I think of why I still own a car, it is simply because I have a space. If I did not, I would have rid myself of a car years ago.

Full article:
http://meanstheworld.co/community/still-car
 
Last edited:
When I went to BU, there was a Prius outside my apartment that was parked around December and didnt move until April. Not once was the snow cleared off. It even had 4 tickets, I assume because the stickers were never visible.

As long as parking is free, theres no reason NOT to have a car. JUST IN CASE.

And the problem with just in case is that you use it more. Its 20 degrees out today. Might as well use the car to go to Shaws, its cold. Its raining, Ill take the car to class. Etc etc.

No car, those trips still happen, but if its free and the car is there....then why not use it?
 
When I went to BU, there was a Prius outside my apartment that was parked around December and didnt move until April. Not once was the snow cleared off. It even had 4 tickets, I assume because the stickers were never visible.

As long as parking is free, theres no reason NOT to have a car. JUST IN CASE.

And the problem with just in case is that you use it more. Its 20 degrees out today. Might as well use the car to go to Shaws, its cold. Its raining, Ill take the car to class. Etc etc.

No car, those trips still happen, but if its free and the car is there....then why not use it?

Sounds like the old adage - If the only tool in your toolbox is a hammer....
 
^ good to see this out there. Seems to me a per-neighborhood auction is the way to settle this, with prices rising (or re-set annually) to a market clearing price (where the number of permits issued equals the number of spots. Rich areas would pay more, easing the "regressive" aspects.

Unit owners get a tax rebate equal to half a car, or a whole Hubway membership, and the rest should go to amenities like heated bus shelters (winter) and bike share (non-winter)

This is where we need to be. Unfortunately, no politician appears to be in favor of anything like this at all. Plus the city seems utterly unwilling to count the number of spaces.

It would also make sense to have tiers of parking to match the requirements of the neighborhood. Offer a 24/7 permit and a night/weekend or day permit with varying price points. The city could then designate spaces to match demand and open up more resident spaces to business use (and meter revenue) during the day.

With a tax rebate or recycling of funds into community projects this is a huge benefit for everyone. Don't have a car? Here's some money! Have a car? Now you can move your car without fear of losing your space forever. Completely illogical it hasn't occurred yet.
 
When I went to BU, there was a Prius outside my apartment that was parked around December and didnt move until April. Not once was the snow cleared off. It even had 4 tickets, I assume because the stickers were never visible.

As long as parking is free, theres no reason NOT to have a car. JUST IN CASE.

And the problem with just in case is that you use it more. Its 20 degrees out today. Might as well use the car to go to Shaws, its cold. Its raining, Ill take the car to class. Etc etc.

No car, those trips still happen, but if its free and the car is there....then why not use it?

Forget the viability of building a self-driving car for a moment. Do you agree that the existence of a self-driving car sharing service would make people less likely to own a car?

It's cold, it's raining, and I need groceries. Would I rather pay to own a car year round with all of its expenses or would I prefer to take 30 seconds on my phone to summon an affordable (no driver to pay/tip/put up with) car to throw my groceries into and take me home (while I watch Game of Thrones on the video board)?

Yes, free parking makes people more likely to own cars, and there should be a hell of a lot less of it. But there is a tipping point for each person, and I think self-driving cars would greatly reduce car ownership in cities.
 
Boston's resident parking policy seems to be designed more out of bureaucratic laziness (we don't want to count all those spaces/manage tiered parking zones) than out of actual coherent policy.
 
Forget the viability of building a self-driving car for a moment. Do you agree that the existence of a self-driving car sharing service would make people less likely to own a car?

It's cold, it's raining, and I need groceries. Would I rather pay to own a car year round with all of its expenses or would I prefer to take 30 seconds on my phone to summon an affordable (no driver to pay/tip/put up with) car to throw my groceries into and take me home (while I watch Game of Thrones on the video board)?

Yes, free parking makes people more likely to own cars, and there should be a hell of a lot less of it. But there is a tipping point for each person, and I think self-driving cars would greatly reduce car ownership in cities.

It will reduce car ownership but potentially increase car trips, and with it, congestion.
 
Well, replacing personal auto with shared auto won't increase number of car trips. But supposing that more people who don't have cars will start taking more car trips, perhaps that will increase # of trips. On the other hand, if people who now own cars then sell them, they will end up taking fewer trips because they're paying per trip.

Anyway, all this is why you also apply the same technology to driverless buses that run every few minutes. And that's possibly an easier problem since they follow predictable routes.
 
It will reduce car ownership but potentially increase car trips, and with it, congestion.
I don't worry about congestion due to increased shared use much because shared-autos-without-parking will be much less congesting than personal-autos-with-parking for at least 3 reasons:

- Shared autos will probably figure out how to encourage true ride-sharing at congested times (like a jitney or sharing a cab)
- When parking is no longer essential to car trips, we can return returning street space to use in "moving" rather than "storing" (probably making walking and biking way more convenient such that car trips don't skyrocket, or yes, turning them in to auto lanes)
- eliminating the traffic of "circling the block" looking for parking (which is particularly congesting because it is distracted and stops unpredictably)

Moving people is a *very* high-value city activity that's how we know that parking is too cheap: its been allowed to crowd out actually moving people.
 
When I went to BU, there was a Prius outside my apartment that was parked around December and didnt move until April. Not once was the snow cleared off. It even had 4 tickets, I assume because the stickers were never visible.

As long as parking is free, theres no reason NOT to have a car. JUST IN CASE.

And the problem with just in case is that you use it more. Its 20 degrees out today. Might as well use the car to go to Shaws, its cold. Its raining, Ill take the car to class. Etc etc.

No car, those trips still happen, but if its free and the car is there....then why not use it?

I believe after 3 days a car is considered abandoned on a public way in Boston. Should have reported it so they could tow it away.
 

Back
Top