"Cars are big computers, and have been for a while"

Doesn't mean we need to treat them like software development platforms. Most of the challenges are still at OSI layer 1 and most of the problems to deal with are mechanical: Tire still wear out, disc brakes still function much as they did 10 or 15 years ago (and wear out), wheel bearings, hydraulic systems for power steering, etc.

Given that cars have much longer lifespans than typical consumer electronics, the last thing I want is a car with a built in 10" touchscreen running some ancient version of Android. I'd rather have some sort of standardized modular system where I can put my own tablet-sized computer in an appropriate place and have it get power, audio sync with the car's speakers. Look at the crap infotainment systems built into 2015 and 2016 model Toyotas, Hondas and Nissans and compare it with a good $400 Android 6.x tablet. Do you really want that same un-updated OS and software for ten years?

The idea of the "double DIN" size car stereo mounting location was a good one, but we need something bigger than this, and with standardized cable connectors:

http://www.autodvdgps.com/ebay/ZC620/ebay-3.jpg

Carmakers, please find a way to separate the 'infotainment' part of the car's computer from the embedded computer that's actually required for the car to run (engine timing/valves, fuel injectors, mixture, emissions monitoring, intake oxygen sensor, etc).

Oh my gosh yes!

The car should be the peripheral, not the main compute device.

Is it a big deal to plug in a 10-year old USB mouse into your computer? No, because that is a well-defined interface. Ditto for some other peripherals. This is what I'd like my car to be.

I don't think we'll see that anytime soon though. Unless everyone gets on-board with a Bluetooth standard that exposes sufficient functions to the phone/tablet to be useful.

But you can also look at this and the point of view of carmakers. What an amazing boon to sales. This is a way to finally solve the ultimate problem of selling cars today. A car you make today has to compete with the extremely good car you made five years ago that still works great, still looks great, and is much much cheaper.

Basically an RPi3 / RPi Zero with a single connector and a number of well defined interfaces for audio, video, talking to car systems, networking, etc...

The possibility to swap I/O devices would be nice too, so you can integrate stuff like Soli (Google ATAP) when modules for that shows up.

And I want that same thing for smart TV:s too. Ideally both of those could use the exact same standard too!

Sticker Price I found was $12,825 ,so let's go with that. Cost to Own for 5 years I found is $28,855. I assume this includes depreciation and that the car fully depreciates in 5 years. If the car lasts for 2 years, the monthly cost comes down to $796/month. If the car lasts for 5, the monthly cost is $476/month, which is a little more than half.

Now this is the most budget new car I could find. Straight up doubles in cost. Or you get the situation where people just decide to drive with the legacy version anyways, which sounds a lot like a literal ton of Windows XP at 50mph. And all because you couldn't come up with a universal plug for car entertainment systems.

It will have absolutely zero impact on their bottom line sales to stop updating their carOS, since they have been doing it for years already and have never done it in the past. Every car with bluetooth / cellular / wifi support has to, by design, contain a computer to provide that service. Every car with a dash more sophisticated than dials has a computer in it.

Nobody has ever considered the security or ethical consequences of these wholly proprietary computers in their cars, and in the same way modern John Deere tractors are complete black boxes with no ability for quote unquote "owners" to inspect, audit, modify, or replace the software part cars have been the same way for years, even if the computer hasn't given you an obvious touchpad UI to notice it.

On the other hand, things with engines can hurt people so you cannot allow unconstrained access to that port. Safety, IP and "who owns what" will make things more interesting as things emerge.

I did pretty much this in my car, and it was surprisingly easy. ~7 inch tablets are just about the same size as Double DIN radios, so a little but of modification to the mounting bracket, some wiring, and a used Nexus 7 I picked up on ebay got me this: http://paste.click/RILfPr

It's hooked into the car's speakers via a USB dac and a small amp, and pulls sensor data from the OBDII port with a cheap bluetooth scantool, but otherwise is completely independent of the car's systems. Turns on and off instantly with the car's power thanks to Timur's Kernel (super low power standby really, but the battery drains isn't much different than having it completely off).

I also attached a wireless remote via OTG so I don't have to mess with the touchscreen for simple things like volume and media controls while driving.

So it's definitely a viable option. I'm surprised no one has come out with a standardized D-DIN to <popular tablet> bracket yet.

One of the nice things about Tesla is their regular, free software updates. Even the earliest Tesla Model S models can be updated to the latest software. You get all the features that the 2016 Teslas have, except of course where the required hardware is not present (like Autopilot).

https://upload.wikimedia.org/wikipedia/commons/d/db/Vfd_car....

Those are vacuum fluorescent displays and not LEDs, and were used in new vehicles (at least by Ford) until the early 2000s. They are still a great choice for vehicle instrument panels and are commonly used for aftermarket retrofit units (see for example http://www.parrautomotive.com/index.cfm/page/ptype=product/p...)

I've been really enjoying my Mazda 3's setup though; it's got the touchscreen thing, but it actually turns the touchscreen off while you're driving, and has a little dial near the console that you can move around like a joystick to manipulate the controls. This sounds weird, but is highly intuitive, and once I knew my way around the controls I could easily work the display blind, with my eyes still firmly on the road in front of me. I'm OK with a compromise like this. The system isn't perfect, but the input method seems like it achieves the best of both worlds.

My 2002 Nissan has physical buttons and also little feeler dots to enable completely blind operation, and the information is displayed high on the dash so you don't have to look far.

But by far the best experience I have had with in car entertainment was a simple bluetooth link to my phone and a physical volume knob on the head unit.

Hop in, start driving and the music is automatically playing what I was listening to wherever I was before. Volume up or down as appropriate, eyes never leaving the road.

If your infotainment system is down but there is nothing physically wrong with the AC compressor or fan system it should not prevent you from using the AC in 43C weather. That's as close as I've seen yet to a touchscreen-system software problem breaking the mechanical driving functionality of a car.

Regular, free software updates are nice as long as the regular, free software updates are nice.

As soon as whoever controls the software updates decides to do something not nice, however, those updates become a huge liability. It's annoying enough with things like evergreen browsers or Windows 10, all of which apparently break basic functionality all the time. But with cars, where an enforced OTA software downgrade can reduce what you bought to something less than you thought you were paying for, and you can have the original back for a few thousand bucks? That's just a legalised extortion mechanism IMNSHO, and I don't want anything to do with it.

In the same vein if you're buying a relatively new car you're going to have an unknown amount of reliability based on new changes to that model year. How the company deals with that down the road is a similar issue. I remember looking into a Hyundai a long time ago only to find out that they were denying warranty repairs on clutches with ~4k miles that hadn't been abused but were failing.

Interesting. I'm curious about how Tesla will do given their very different commercial and technological models. However, I've never driven one, and the anecdotal comments about them that I've noticed have given me an overwhelmingly negative impression of their brand. Maybe what I've seen so far was just unfortunate and not a fair representation?

I imagine the updates will stop for older models about the time they decide that supporting that version of whatever hardware costs more than they get out of doing it.

Lets see if they are shipping high volumes in 2020 and still supporting 2012 cars.

Eg.

When the parking break was on and you forgot and tried to drive, the 2003 had some resistance. On the 2014, it seems that it disengages the brake for you if you are not in Park. Which can be annoying if you have to stop on an incline with traffic.

It also seems to regulate your speed in very wet conditions. On a straight road, in moderate rain, it definitely feels like I need to push the accelerator down more to get up to speed.

I admit the 2014 model is nicer to drive, but I don't know what will happen if the electronics get fried.

I.e. not a software "write enable" switch.

The solution to this issue isn't treating the car manufacturer as a hostile party and trying weird tricks to circumvent their future actions. The solution is ensuring that the car manufacturer isn't hostile in the first place.

Ideally I'd like to do that with market forces, but that requires the average consumer to understand a lot more about the technology and its implications than is realistic so they can protect themselves.

A more realistic alternative is probably statutory regulation and penalties for non-compliance that can become an existential threat to the manufacturer's business.

Completely off-topic, but if I saw a new truck with the old-style floor-mounted stomp pedal for the dimmer switch, it'd almost be a shut-up-and-take-my-money situation. I cannot comprehend why automakers keep piling more stuff into the steering wheel. Some of the little sedans have handles and buttons coming off them every which way.

Cars are so much more reliable than they were even 25 years ago. Reliable and longer lasting. My first car broke down if you looked at it wrong. My current car does not have that problem even though it is just as old/has just as many miles on it than my first car did at the time. Breaking down on the side of the road used to be a common part of my life, now it hasn't happened to me in the last 8 years.

edit

I also find AAA much better than crawling under my car in the middle of the night 50 miles away from anything...in general. But I know that's a personal preference.

I hate that cars are bloated with superfluous components that always break before the car itself does. That said, there are some gadgets that I would miss on a fully mechanical car. I'd be happy to find a manufacturer that could balance that correctly.

Low tire pressure sensors are required in all new cars after 2011 for safety as driving on under inflated tires is very dangerous. It's a much needed safety feature. Your car is going to get more "bloated" every year as new safety features are required each year because they work at reducing crashes.

I am not sure who just blindly fills up their tires without checking the psi as they go.

The break in alarm sometimes starts to blink the lights on its own. You have to do some "open the doors, lock the doors, open them again" sequence to fix it. It took me year to figure out what's wrong and how to do the right moves.

The radio doesn't work anymore, as there is some kind of superuser code that's required every time the battery dies.

The headlights have some computer control. They point to wrong altitude.

Mechanical problems happen too. But they often happen by wear. You get significant warning before it's completely kaput. And you can often guess what's wrong from the symptoms quite easily, because the erratic behavior is more consistent.

I've tried to get the dealership to disable the door auto-locking (get above 5 mph, doors lock) "feature" half a dozen times, and I've given up, since they disable it in the software and it's turned back on before I leave the parking lot.

ABS brakes are awful, and down-right dangerous if you are not expecting their flaky behavior, and rather want them to perform in a predictable and consistent way. It's terrifying driving an ABS vehicle on icy roads.

Who hasn't had a largely irrelevant, but very expensive, sensor burn out and trip a buggy corner case in the engine firmware?

I realize I sound like a Luddite, but my favorite vehicle so far is still my first, which was a bare-bones, rusty 1985 F-150 with a four-speed manual and a straight six. It's still on the road, with over 300,000 miles on it, although it has become something of a truck of Theseus, aside from the frame and tranny. You can tear the whole thing down and put it back together with a pretty basic set of mechanics tools. It wasn't perfect - starters and solenoids tended to go bad, and mileage was a tad worse than the current models - but it's simple, so that when things go sideways, you patch up the one component that's broken, and trundle onward.

Where are you going to get the parts if you are 50 miles away from anything?

There's definitely something to be said for the maintainability of a vehicle where a small tool box worth of tools is enough to fix 90% of all possible problems.

I recently got a Pioneer 4100 NEX aftermarket CarPlay device fitted in my 2010 Camry just for this. Unfortunately, CarPlay itself has many usability issues that have not allowed me to use the system the way I had imagined.

There is actually a way to use the native Android interface if you have a rooted device: http://forum.xda-developers.com/showthread.php?t=2451980

I do not use this personally, but I was looking into the Pioneer 4100 NEX myself and plan to buy one after finding this workaround.

They have, sorta (but it's still rather limited in functionality):

https://www.android.com/auto/

Android Auto and I believe Apple Car Play both run totally from your phone/tablet, with the car just being a view screen. Thus, you get updates when your phone updates.

Maybe a switch to flip the display to access direct car controls (but really just give me hard-buttons/dials etc for climate control and keep the consumer electronic part simple and separate).

Extra points if there's a standard for sending car control api calls back and forth to the phone, but 90% of utility could just be handled with a remote touch display.

Use a standard modular platform so you can hook up whatever SoC you like. Whenever the old one gets dated, upgrade it and get all the latest fancy radios.

The cumudgeonly John Muir ( who wrote "Volkswagen Maintenance for the Compleat Idiot") is on record as saying cars would be safest if "... drivers were strapped to the front like an Aztec sacrifice" (paraphrased). Exploits to take over cars through the infotainment systems only reinforce my prejudice.

Actually, regenerative breaking means they last quite a bit longer.

IIRC, a CAN bus design is essentially a jazz model for connecting different subsystems with no built-in security standards. For some reason, I'm not sure externally connected hardware, OTA-updated software, a vehicle's essential mechanical control systems, and implicitly trusted access from everywhere to everything via a centralised bus is a good combination.

Some of the relevant technology is patent-encumbered as well, which is rarely an ideal basis for standardisation and interoperability IMHO.

The only thing that still worked was the engine management computer (it still ran fine) and the headlights which seemed to be wired directly. I think the brake lights also worked but I'm not sure.

Tesla may be all cool and willing to update software, but it's very easy to be that when your fleet is all shiny. Come back when someone finds a security flaw in the 25 year-old model that is now in the hands a highschool kid as his first car. Come back when the 4g antenna in the dashboard can no longer communicate over the now-standard 67gxrr mobile network. That's why there are warehouses stacked full of parts all over the country. That's the car biz.

If I didn't have that know-how (or a friend or family member who could help me out), I'd be SOL. And this is hardly a one-off thing, and hardly specific to me. Plenty of people are only able to afford cars if they can fix them themselves.

Moreover, I hope to keep these cars running 10 or 20 years further down the road. The only reason I can hope to do this is because of aftermarket manufacturers and no technical restrictions on what I can replace, upgrade, or re-make on my cars.

A friend of mine has had weird problems on his relatively modern car because of what's probably a cold solder in the fuse box on his car. That fuse box costs $700 to replace PLUS whatever the dealer is going to charge you to reprogram the computer in it. Why does a fuse box need a computer in it? Hell if I know. How do you fix this once the dealer no longer keeps around the cables and software to reprogram that computer? Either replace every single computer in the car with computers from a car with a working fuse box, or reverse-engineer the proprietary protocol the fuse box talks and re-engineer a replacement.

Oh, except that latter choice is legally questionable, thanks to the DMCA.

I expect a lot more cars from this era to end up as fancy bricks.

That's a damn shame; there's no reason to discard a nearly-functional pile of parts just because some auto manufacturer is too stingy to help you fix, or not hinder you from fixing, the thing they built and sold you. I recall a study that showed that keeping an old car functioning was just as efficient pollution-wise than building a whole new car to replace it with.

I guess the upside here is that there's going to be a decent amount of demand for folks like me who have both mechanical and electrical chops in the car repair scene in a few years.

This is the sort of thing economies of scale bring. It's evidenced by CPU and GPU binning, whereby it's cheaper for the manufacturer to make one part and segment the market with simple switches to disable functionality (in some cases, disabling parts of the chip with flaws which render it otherwise unsellable).

Car makers have been doing this for ages with simpler features, such as including air conditioning or intermittent wipers but requiring the user to pay exorbitant amounts for the switch on the dash to activate them.

not quite. it's possible you can't afford to make the fully-unlocked version at scale unless you can charge extra for it.

A 75kWh battery pack that is limited to 60kWh is better than having a real 60kWh pack.

The extra over-capacity means that it won't be charging to it's true 100% capacity, and not discharging to 0%. That puts less wear on the cells, and means faster charging. (The sweet spot for charging is in the middle of the battery's capacity - charging speed is slower if the battery is near 0% or near 100%.)

This is the first thing I've heard coming out of Tesla's court that really put a bad taste in my mouth.

Tesla cars will depreciate more slowly, with resale prices closer to original MSRP. That's all good news for their leasing program, their resale value guarantee program, and their certified pre-owned program.

There are a lot of upsides to reducing battery wear. Battery wear is a big reason my 2012 Nissan Leaf is worth less than 1/3rd of its original MSRP.

If I were in the market for a car right now I'd be looking for about a 2008 model year. All the significant depreciation has already been realized and you can find some real values in good condition with a lot of life left if you shop around a bit.

Why? Wouldn't the weight reduction be useful?

> If you look at this as a $6000 or so discount on a new Tesla, it actually seems like a pretty good deal.

Yes.. but the whole battery is still being used, and is still in a position to require service. It could be the case that I have a 60kW tesla, but still need to pay the full 75kW repair price, because of the way they structured this.

Servicing costs aren't really something you need to worry about, though. Tesla battery packs are warrantied for 8 years. After that time, battery costs will have greatly reduced and there may even be cheap aftermarket replacement packs available.

These are systems designed to prevent accidents in case the driver misses something, but NOT for daily use!

But you know people _will_. That's what happens with affordances. Design. It changes how people interact with things, and not always in the intended way. It's _very_ important to consider when you're talking about the operation of machines that can kill.

Indeed, the presence of emergency brakes could very well reduce the number of accidents, while at the same time making people rather lazy and eventually _causing_ some. Like taking bigger risks on your bicycle when you're wearing a helmet.

Is there a psychological term for this thought process?

The thought process: People killing people is "okay" but people killing (overall) fewer people because they are relying on "emergency auto-brake systems" is "not okay".

For me, it is a no brainer to reduce human casualties as much as possible. But it seems others disagree with that if some automated system is responsible for deaths instead of a human. Is it because blame is harder to place?

I suspect that's part of it. Another possibility is that in some cases, these systems may outperform an average driver but not a good one. Removing control or judgement from a good driver in favour of the automated system is then an unnecessary loss.

Of course, you also have the problem that around 100% of drivers think they are better than average. The above is a reasonable and valid concern for drivers who actually are better than the alternative automated system, but there's also a problem with other drivers who mistakenly believe they are that good and that the automated system will therefore make them and their passengers less safe.

People are only as careful as they think they need to be, because they're lazy.

This causes your behavior baseline to change as a result of new safety measures, perhaps too far in the wrong direction.

Found this - procedural drift (aviation): http://www.cognitivesystemsdesign.net/Papers/Johnston%20(200...

Take any technology that automates a task. Say it has a 0.00005% failure rate which results in death of a random passerby at the time it occurs. Then take any human being in charge of that same task. Say the risk is 0.05% human error which results in death of a random passerby at the time it occurs.

There are people who will argue that the 0.05% is acceptable and will vehemently disagree with and prevent any attempt to switch to the 0.00005% failure rate because the machine killing someone is unacceptable while a human killing 1,000-fold more people is somehow acceptable. Even if human laziness of an assisted (not completely automated) version of the same task had a 0.005% risk it would still be unacceptable when compared with the 100% human-error rate.

Morally, I can't wrap my head around the reasoning. More deaths seems to be acceptable as long as some person can be "at fault". The moment the death is "blameless" it cannot be allowed.

In other words: 10,000 "someone can be blamed" deaths is better than 10 "who is to blame?" deaths.

I think a charitable interpretation of that driver's actions would be that they expected the car to act like it was adaptive cruising, not that they expected it to abruptly brake for them.

In "autopilot" mode, the automatic cruise control prevents you from hitting the car ahead at all. Touch the brake, and you've disabled both Autopilot and Automatic Emergency Braking. Yesterday, that caused another accident.[2]

If you drive a Tesla, failure to carefully read firmware release notes can kill you.

Compare BMW.[3]

[1] https://www.teslamotors.com/sites/default/files/tesla_model_... [2] http://arstechnica.com/cars/2016/05/another-driver-says-tesl... [3] http://www.theverge.com/2015/1/5/7494989/bmw-i3-self-parking...

(I'm not sure why Tesla would design it that way. While there's certainly an argument for autopilot's steering and acceleration to always be overridable, I don't see a reason why emergency braking shouldn't always be active if it means the car will avoid a collision - even if the driver is fighting against it!)

Is that really true? Are the batteries otherwise identical and you can get stranded with plenty of reserve power?

But as for the second half.. I feel that cars are going to adapt to us, rather than us adapting to cars. It's already happening, as the Volvo steering wheel example describes.

Take a few hundred thousand, buy a collection of low mileage pickup trucks and Honda Accords and store them for 10 years. I bet you'd 10x your investment.

again, just a theory. may be more complex than this.

Not even hydraulics?

Note also this was a light vehicle; turning the wheels was quite hard unless you were moving a bit, and the brakes required a tremendous amount of force, it took me a long while to not hit contemporary '70s power brakes too hard.

What else: the windshield wipers were powered by engine vacuum, and slowed down when you accelerated (this one had a V6, so it was pretty powerful, and pretty zippy 0-30 MPH, only 3 gears forward so it lagged after that). Oh, yeah, see https://news.ycombinator.com/item?id=11919653 cars thought that era had electro-mechanical turn signals. In the case of this Jeep, the turn signal control was kludged onto the steering wheel, there was a little rubber wheel which contacted the steering wheel shaft to disengage it after you'd turned.

And not entirely reliable, 10-12 years old when I drove it, it had a towing hitch on the front, which was used more than once.

IIRC, it's been said by Cory Doctorow that a pacemaker is a computer I put into my body, and a car is a computer I put my body into.