14 ms·

The whole thing reeks of NIH and some poor saps trying to save their butts by trying to be forcefully relevant. Apple and Google already solved the UI problem

The whole thing reeks of NIH and some poor saps trying to save their butts by trying to be forcefully relevant. Apple and Google already solved the UI problem - you don't need to have the damned ugly in-car UI anymore. Just stick decent hardware with Android Auto and CarPlay support and be done with it. Make decent looking apps for each platform to provide functionality outside of platform supported features. I just don't understand why on earth would Ford need to have their own system with QNX, WiFi updates and Maps that are not free. What problem does it solve that is worth solving at the expense of reinventing the wheel and making users deal with yet another system? Edit: One way Ford to justify it is to think of the non-SmartPhone, non-Apple, non-Google users. Though in terms of numbers, at least in the US this bunch is a tiny minority. Ford could give them a Moto G free with every Car for not having to pay for QNX licensing fees, maps and updates etc.! OR they could limit the QNX based independent system to lower end cars and make the higher end ones compatible with Android Auto and CarPlay. There's also the question of longevity - what if in 2017 Apple and Google stop being compatible with the older version of Carplay/Auto compatible implementation Ford shipped in their 2014 car? What if users switched to non-Apple, non-Android phones? In that case it would make the in-car system useless for the user. So given people use cars for way longer than SmartPhones - I guess it makes some sense for Ford to have an independent system. Regardless they should still rid themselves of the NIH and stick with Android based UI and not have their own based on QNX. That part hardly makes sense.

mtmail 12y ago
Car companies probably don't want a vendor lock-in. Android Auto and CarPlay may be two options but what if one goes away or becomes unattractive in the future?
- blinkingled 12y ago
  Great point, I realized soon after I posted and so I updated my OP. But still they could have gone with Android based system instead of QNX - it's open source, the Lollipop UI is gorgeous and since it would be a fork, there's no lock-in worry. Developers would be able to write / port apps and extensions easily as well.
  - angersock 12y ago
    I can't put this delicately. It's a car. It's not going to have an app store. You don't want an app store. You don't need an app store. QNX is a well-understood RTOS--look up "hard real time". There's a lot more important things in a car for an OS than being able to display a notification on Facebook; things like deploying airbags, coordinating brakes, and all these other things.
    - maxerickson 12y ago
      The Sync system isn't doing any of the important things you mention. It's infotainment.
      - sjwright 12y ago
        Furthermore the notion that you'd be sharing core processing between safety-critical car functionality with infotainment functionality is downright scary. These should always be (are they?) discrete systems. I can cause my BMW's i-drive (CIC-HIGH NBT F20 2013) to completely crash if I perform a specific action. After fifteen seconds the system detects the freeze and reboots. The way the reboot occurs strongly suggests the reboot is performed by a hardware level supervisor. But the rest of the car continues to function.
  - Matthias247 12y ago
    CarPlay and Android Car don't help you very much: They require a phone to operate. If you don't have one you get nothing. That doesn't sound like a big issue now but keep in mind that car multimedia is in development for about 4 years and then the cars are in the market >= 15years. This means if you now start to design a system around CarPlay you would need a iphone that supports the current CarPlay protocol in 2034. That's very unlikely. And when you don't have one you get nothing at all, neither music nor navigation. Directly building a system on top of android instead of QNX or a custom linux distribution is certainly possible. Every approach here has some pros and cons. Keep in mind that not all of android is open source and available for everyone. You can't directly use a normal Android phone UI because it is not certified for in-car use (driver distraction regulations). But you could at least use it's UI framework and some of the widgets. For the widgets there is the question if you want them. If they started development 3-4 years ago and based on android you would now get a brand new Ford with a Android 2.3 UI. Wouldn't look sexy either and even less in some years from now. Therefore going with some custom UI can make sense.
- function_seven 12y ago
  Can't answer your question for CarPlay, but wouldn't Android Auto forkable if need-be? Not desireable, but an option for a car company wishing to hedge its bets.
toomuchtodo 12y ago
Tesla rolled their own Linux-based in-car system, and its absolutely gorgeous and a pleasure to use. I highly recommend stopping by a showroom to see it in person. > I just don't understand why on earth would Ford need to have their own system with QNX, WiFi updates and Maps that are not free. I agree this is the wrong way to go about this, but rolling your own solution should not be indicative of NIH syndrome if existing options are sub-optimal.
- r00fus 12y ago
  The also lured/poached some UI folks from Apple [1] to work on it. You think risk-averse Ford would attract the same talent as a hot-as-shit disruptor with a large upside stock potential and the chance to "redefine the world"? [1] http://9to5mac.com/2014/07/29/much-of-teslas-ui-design-team-comes-from-apple-video/ http://9to5mac.com/2014/07/29/much-of-teslas-ui-design-team-...
smackfu 12y ago
>Apple and Google already solved the UI problem - you don't need to have the damned ugly in-car UI anymore. I haven't seen a good example of those systems not sucking yet. Also, they require that you plug in your phone, which is really a downgrade from our current Bluetooth systems. >Ford could give them a Moto G free with every Car for not having to pay for QNX licensing fees, maps and updates etc.! Or they could just build the Moto G into the car! And now you are back to square one.
- mikeash 12y ago
  Are they seriously not wireless? What the hell were they thinking?
emcrazyone 12y ago
I was one of the architects working on Ford SYNC. Internally, Ford has a big problem with open source software from a legal stand point. You have a fortune-5 company here. Translation for you: They are a very large target for lawsuits. The open source software provides no indemnification. You also have to remember when SYNC was in the planning stages long before the public knew about it. I was part of the original team responsible for taking the whole concept to reality. I can tell you that internally, it was one of the most well run software architecture teams I was ever apart of. The middle level managers really know how to run software teams which is not something you might expect from a big old automotive company. But in many ways, our hands were tied. Recall about the time Bill Gates and Bill Ford riding around in a model-T. A lot of us didn't want Windows for automotive and were trying to champion Linux. Also, the iMX3 which was in the first generation SYNC modules was at the time a slow processor. And to top things off, a fair bit of the software running on the SYNC module is not written by Ford. Ford partnered up with a crappy company called B-Squared. The Maps are not Ford's either. When I was on the team, Ford had partnered with INRIX for Maps, traffic and direction. SYNC is just not a simple single board computer with some apps running on it. There is an entire eco-system build around the SYNC module because it's connected to the vehicle networks (CAN Bus). Microsoft & Ford being partners (Windows servers in data center), Microsoft wanted to handle some of the software development but when Ford asked Microsoft to sign off on some legal agreement asking MS to take any and all responsibility for things like inadvertently deploying an air bag, they backed off. Also, Ford SYNC as I said has an eco system of roughly 20-30 backend applications that support it. Some of these have to do with 911 assist and the TREAD act. The TREAD act is the US Federal Government's oversight on safety claims. Because via SYNC you can report problems and have your vehicle serviced, their is a lot internal logic that keeps track of what is called EOL (End of Line) data about a vehicle which needs to funnel into TREAD act reporting. Think about the not so long ago Toyota problem where they tried to blame floor mats for gas peddles getting stuck. EDIT: Also, the GUI is not Microsoft CE native. The WindowsCE is just the OS running on the SYNC module. The GUI was Adobe ActionScript. Trust me, many of us yelled very loud about how stupid an idea this was but because B-Squared own the implementation (not the architecture), they were aloud to chose whatever they wanted to meet Ford specs. The results are a history less: JD Powers gave it SYNC poor ratings and many of us (myself included) got the hell out of dodge. Middle managers, managing the day to day software architecture and specifications, were caught between a rock and a hard place with senior management (Allan Mullaly, Mark Fields, Bill Ford, Marcy Klevorn) pushing these partnership relationships we were forced to work within.
roc 12y ago
> "Apple and Google already solved the UI problem" Have you seen any demos of these systems? Because I thought this as well until the video evidence began to mount. Both CarPlay and Android Auto are slow, laggy and overly complicated solutions. And that's before we account for both systems' odd reliance on cloud-driven voice navigation. And that's without trying to account for the frankly-unlikely notion that people will fish their device out of their pocket/bag, and physically plug it into their car, with any regularity. Both Apple and Google's efforts are attempts to cram their existing platforms into a use-case where it isn't needed and it just doesn't fit. It's akin to Microsoft trying to cram kbm windows onto a tablet. That is not the way forward for car manufacturers. That's not the way forward for anyone. The primary desired integrations by the user are: being able to listen to the audio that's on their device and exchanging voice with contacts on their device. Even turn-by-turn is a thing that most people, most of the time, don't need. And when they do need it, the odds of having their own car, or a car in which they had any choice on the underlying in-dash technology, are low. Niche groups might select a car/radio and regularly plug their device in to attain these features -- just as niche people suffered various iterations of Windows Tablets over the last decade. But there will be no widespread adoption or use. What's needed is an update to bluetooth [1] and dash designs that acknowledge mobile devices by providing a universal mount and some signalling method that can shift those devices, themselves, into a "car mode". So 80% of the time, 80% of people aren't mounting anything and are having a better experience. And when they do want some device feature, they don't need to hunt through the purchase process for various possible mounts, suffer fiddly installs and then live with obscured dash controls. So I see every reason in the world why Ford would need a competent in-dash solution above and beyond simply supporting CarPlay and Android Auto. As to why Ford would want a QNX solution with wi-fi, for-pay turn-by-turn, etc? Fleet sales / rental services. [1] improve the audio handling. improve the pairing situation. voice command pass through.
- gmarx 12y ago
  I'm probably one of the niche guys but I see wider applications for this. I like in dash GPS which I have from an after market unit. I use it frequently. The problem is that to update the maps, you need to purchase a CDROM for a hundred smackers. Once you have it for free on your phone you are not going to do that. So I use the phone. Hooked up, the voice instructions play through the stereo. Direct connection is noticeably superior to blutooth for this purpose. The only problem is the awkwardness of screen control access. I could just mount the phone to the dash but the iPhone screen is a little small. What I want is to connect my phone and then have a nice big touch screen on the dash that functions like the phone. I don't care much about voice control (I don't have Siri) but I suppose it would be nice. I'll look for the videos. It seems improbably to me that they could screw this up but I am often dissappointed
mrpippy 12y ago
> Make decent looking apps for each platform to provide functionality outside of platform supported features. CarPlay and Android Auto both do not allow arbitrary apps (which would be needed to i.e. control the head unit's FM/AM radio or climate controls). I expect this will happen someday, but it's not here yet.
Lagged2Death 12y ago
...non-SmartPhone, non-Apple, non-Google users ... at least in the US this bunch is a tiny minority. "Two-thirds of Americans now have smartphones:" http://www.engadget.com/2014/02/11/two-thirds-of-americans-now-have-smartphones/ http://www.engadget.com/2014/02/11/two-thirds-of-americans-n... One-third is not a tiny minority.
- matthewmacleod 12y ago
  I suspect it's a tiny minority of the people who are in the market for cars with built-in interactive systems like this, though.
- umsm 12y ago
  What percentage of new car buyers have smartphones? I have a feeling that most people without smartphones will not be in the market for a new car...
  - wcfields 12y ago
    Even if it was 99.99999% I would bet that there's still a huge percentage that has no idea what their phone can do, let alone Bluetooth pairing it. I'd even guess that many people think Bluetooth is the name just for those douchey headsets.
Animats 12y ago
There's also the question of longevity That's one of the big arguments for QNX. The QNX microkernel is very stable. It barely changes from year to year. Versions from 15 years ago run fine. There's no need for OS updates. Cars last a long time. I have a 1985 Ford Bronco in my driveway, and it still has the original EEC IV electronic engine control unit, still working fine. That was designed for a 30 year lifespan, and it made it.
- luckydude 12y ago
  I agree. I was friends with Dan Hildebrandt, he was one of the (few) guys who worked on the actual microkernel. In my opinion, QNX is the only example of a viable microkernel OS (it's way the heck better than Mach ever was). I love Linux, use it all day, every day, but I can see the wisdom of choosing QNX. It's very stable. The downside is you don't get all the new shiny stuff that runs on Linux without some amount of effort. I suspect Ford doesn't care, Ford has always been more about "it works" than "it's shiny". Says the dude who just bought a 1994 F15 5 speed so his son can learn how to drive a manual :)
a3n 12y ago
How long is any particular flavor of Android going to be supported, vs QNX or some other unix-like RTOS? Yes, you could contract for longer support, but how much current energy will remain with whoever's engineers you contracted to for supporting that older and older version? Also, QNX is an RTOS. Maybe they actually needed an RTOS. How well does Android do as an RTOS?
PinguTS 12y ago
It looks like you have no glue in terms of software development requirements in the automotive industry. Software development for embedded devices is a totally different story than software development in the IT industry. The is no minimal viable product and then deliver experience with some updates. Everything has to work from day 0. Otherwise you will have a big PR disaster. Mercedes had that with its E-Class model W210. Only a few companies have exemptions from that like Tesla. A well established company like Ford, GM, Volkswagen or so who delivers something like 50.000 cars a month over all their different models requires different software quality requirements than a company like Tesla who delivers 20.000 cars a year and is considered as a Startup, which may have some hick-ups. Another issue is, that you have even in a simple car you have at least 15 different control units and the HMI/MMI is only one of these. Throwing hardware at it means increasing prices, which you don't want as a consumer. Also remind you, that development of a car starts somewhere 3 to 5 years before you buy the car. Because the hardware has to be tested very extensively so that everything works in without problems in most worst environment like from Alaska to Mojave desert. Designing hardware with such a broad requirement is not as easy as you may think. That is the reason why automotive rated components are way more expensive than industrial rated components, not even talking about consumer rated components. The problem of what you call "decent hardware" is, that it is only consumer rated from 0°C (32°F) to 80°C (176°F) vs. automotive rated, which is -20°C (-4°F) to 125°C (257°F). Then there are also EMI requirements and much more. The only consumer company that is close to the quality requirements of the automotive industry is Apple. But Apple has a different advantage over the automotive industry, which is they sell millions of their phone within a month vs. a carmaker sells a million of a model may be within 3 to 5 year, depending on the market level (small car vs. premium car). The iPhone within a car is only one ECU, but a car as at least 15 of them (up to 100 in the premium class, like Mercedes S class, Porsche). So the cost structure is completely different. Another thing: Android Auto and CarPlay are no OS at all. Both are just interface definitions. Both are implemented by QNX for example. That said, QNX was one of the first suppliers who teamed up with Apple to provide an integration of the iPod.
superuser2 12y ago
Do not do this. In 15 years, your car will still be around but there is a very good chance that neither CarPlay nor Android Auto will exist anymore. You're going in the right direction - this is slightly less stupid than having something like a Pandora client permanently installed in your car such that it cannot and will not ever be updated once the next model year comes out and Pandora makes breaking API changes or ceases to exist. The only reasonable thing to do is provide long-lived open standard interfaces like the 3.5mm analog connector. Anything else is deliberately anti-consumer planned obsolescence. It doesn't matter if you think "Android is the platform of the future" or something. It probably isn't. Don't tie something with the release-replacement cycle of a car to a disposable device like a smartphone.
- twelve40 12y ago
  This puts things in perspective, I appreciate that. However, Ford (and others) seem to already tie their cars to one or another type of transient tech. Like the above-mentioned Inrix for maps, or my Pioneer stereo phoning home to some unidentified service to get the traffic. Is Inrix going to be around longer than Google or the 3.5mm jack? I'd rather automakers provided something like a simple, reliable smartphone docking station, with speakers, mics (don't skimp on those like Ford does), and maybe some kind of a display integration, and leave the UI and the smarts to the fast-moving mobile market. But they seem to prefer to force-feed you those crappy, sometimes insanely overpriced systems and web services that still don't seem likely to outlive Google or the car itself.
sandGorgon 12y ago
is this a certification issue ? I know that QNX is certified for several CAN-bus deployments... maybe Android is not certified for those kind of environments. Again not sure if this is part of the vehicle CAN bus anyway.

mtmail 12y ago

Car companies probably don't want a vendor lock-in. Android Auto and CarPlay may be two options but what if one goes away or becomes unattractive in the future?

blinkingled 12y ago
Great point, I realized soon after I posted and so I updated my OP. But still they could have gone with Android based system instead of QNX - it's open source, the Lollipop UI is gorgeous and since it would be a fork, there's no lock-in worry. Developers would be able to write / port apps and extensions easily as well.
- angersock 12y ago
  I can't put this delicately. It's a car. It's not going to have an app store. You don't want an app store. You don't need an app store. QNX is a well-understood RTOS--look up "hard real time". There's a lot more important things in a car for an OS than being able to display a notification on Facebook; things like deploying airbags, coordinating brakes, and all these other things.
  - maxerickson 12y ago
    The Sync system isn't doing any of the important things you mention. It's infotainment.
    - sjwright 12y ago
      Furthermore the notion that you'd be sharing core processing between safety-critical car functionality with infotainment functionality is downright scary. These should always be (are they?) discrete systems. I can cause my BMW's i-drive (CIC-HIGH NBT F20 2013) to completely crash if I perform a specific action. After fifteen seconds the system detects the freeze and reboots. The way the reboot occurs strongly suggests the reboot is performed by a hardware level supervisor. But the rest of the car continues to function.
- Matthias247 12y ago
  CarPlay and Android Car don't help you very much: They require a phone to operate. If you don't have one you get nothing. That doesn't sound like a big issue now but keep in mind that car multimedia is in development for about 4 years and then the cars are in the market >= 15years. This means if you now start to design a system around CarPlay you would need a iphone that supports the current CarPlay protocol in 2034. That's very unlikely. And when you don't have one you get nothing at all, neither music nor navigation. Directly building a system on top of android instead of QNX or a custom linux distribution is certainly possible. Every approach here has some pros and cons. Keep in mind that not all of android is open source and available for everyone. You can't directly use a normal Android phone UI because it is not certified for in-car use (driver distraction regulations). But you could at least use it's UI framework and some of the widgets. For the widgets there is the question if you want them. If they started development 3-4 years ago and based on android you would now get a brand new Ford with a Android 2.3 UI. Wouldn't look sexy either and even less in some years from now. Therefore going with some custom UI can make sense.
function_seven 12y ago
Can't answer your question for CarPlay, but wouldn't Android Auto forkable if need-be? Not desireable, but an option for a car company wishing to hedge its bets.

toomuchtodo 12y ago

Tesla rolled their own Linux-based in-car system, and its absolutely gorgeous and a pleasure to use. I highly recommend stopping by a showroom to see it in person. > I just don't understand why on earth would Ford need to have their own system with QNX, WiFi updates and Maps that are not free. I agree this is the wrong way to go about this, but rolling your own solution should not be indicative of NIH syndrome if existing options are sub-optimal.

r00fus 12y ago
The also lured/poached some UI folks from Apple [1] to work on it. You think risk-averse Ford would attract the same talent as a hot-as-shit disruptor with a large upside stock potential and the chance to "redefine the world"? [1] http://9to5mac.com/2014/07/29/much-of-teslas-ui-design-team-comes-from-apple-video/ http://9to5mac.com/2014/07/29/much-of-teslas-ui-design-team-...

smackfu 12y ago

>Apple and Google already solved the UI problem - you don't need to have the damned ugly in-car UI anymore. I haven't seen a good example of those systems not sucking yet. Also, they require that you plug in your phone, which is really a downgrade from our current Bluetooth systems. >Ford could give them a Moto G free with every Car for not having to pay for QNX licensing fees, maps and updates etc.! Or they could just build the Moto G into the car! And now you are back to square one.

mikeash 12y ago
Are they seriously not wireless? What the hell were they thinking?

emcrazyone 12y ago

I was one of the architects working on Ford SYNC. Internally, Ford has a big problem with open source software from a legal stand point. You have a fortune-5 company here. Translation for you: They are a very large target for lawsuits. The open source software provides no indemnification. You also have to remember when SYNC was in the planning stages long before the public knew about it. I was part of the original team responsible for taking the whole concept to reality. I can tell you that internally, it was one of the most well run software architecture teams I was ever apart of. The middle level managers really know how to run software teams which is not something you might expect from a big old automotive company. But in many ways, our hands were tied. Recall about the time Bill Gates and Bill Ford riding around in a model-T. A lot of us didn't want Windows for automotive and were trying to champion Linux. Also, the iMX3 which was in the first generation SYNC modules was at the time a slow processor. And to top things off, a fair bit of the software running on the SYNC module is not written by Ford. Ford partnered up with a crappy company called B-Squared. The Maps are not Ford's either. When I was on the team, Ford had partnered with INRIX for Maps, traffic and direction. SYNC is just not a simple single board computer with some apps running on it. There is an entire eco-system build around the SYNC module because it's connected to the vehicle networks (CAN Bus). Microsoft & Ford being partners (Windows servers in data center), Microsoft wanted to handle some of the software development but when Ford asked Microsoft to sign off on some legal agreement asking MS to take any and all responsibility for things like inadvertently deploying an air bag, they backed off. Also, Ford SYNC as I said has an eco system of roughly 20-30 backend applications that support it. Some of these have to do with 911 assist and the TREAD act. The TREAD act is the US Federal Government's oversight on safety claims. Because via SYNC you can report problems and have your vehicle serviced, their is a lot internal logic that keeps track of what is called EOL (End of Line) data about a vehicle which needs to funnel into TREAD act reporting. Think about the not so long ago Toyota problem where they tried to blame floor mats for gas peddles getting stuck. EDIT: Also, the GUI is not Microsoft CE native. The WindowsCE is just the OS running on the SYNC module. The GUI was Adobe ActionScript. Trust me, many of us yelled very loud about how stupid an idea this was but because B-Squared own the implementation (not the architecture), they were aloud to chose whatever they wanted to meet Ford specs. The results are a history less: JD Powers gave it SYNC poor ratings and many of us (myself included) got the hell out of dodge. Middle managers, managing the day to day software architecture and specifications, were caught between a rock and a hard place with senior management (Allan Mullaly, Mark Fields, Bill Ford, Marcy Klevorn) pushing these partnership relationships we were forced to work within.

roc 12y ago

> "Apple and Google already solved the UI problem" Have you seen any demos of these systems? Because I thought this as well until the video evidence began to mount. Both CarPlay and Android Auto are slow, laggy and overly complicated solutions. And that's before we account for both systems' odd reliance on cloud-driven voice navigation. And that's without trying to account for the frankly-unlikely notion that people will fish their device out of their pocket/bag, and physically plug it into their car, with any regularity. Both Apple and Google's efforts are attempts to cram their existing platforms into a use-case where it isn't needed and it just doesn't fit. It's akin to Microsoft trying to cram kbm windows onto a tablet. That is not the way forward for car manufacturers. That's not the way forward for anyone. The primary desired integrations by the user are: being able to listen to the audio that's on their device and exchanging voice with contacts on their device. Even turn-by-turn is a thing that most people, most of the time, don't need. And when they do need it, the odds of having their own car, or a car in which they had any choice on the underlying in-dash technology, are low. Niche groups might select a car/radio and regularly plug their device in to attain these features -- just as niche people suffered various iterations of Windows Tablets over the last decade. But there will be no widespread adoption or use. What's needed is an update to bluetooth [1] and dash designs that acknowledge mobile devices by providing a universal mount and some signalling method that can shift those devices, themselves, into a "car mode". So 80% of the time, 80% of people aren't mounting anything and are having a better experience. And when they do want some device feature, they don't need to hunt through the purchase process for various possible mounts, suffer fiddly installs and then live with obscured dash controls. So I see every reason in the world why Ford would need a competent in-dash solution above and beyond simply supporting CarPlay and Android Auto. As to why Ford would want a QNX solution with wi-fi, for-pay turn-by-turn, etc? Fleet sales / rental services. [1] improve the audio handling. improve the pairing situation. voice command pass through.

gmarx 12y ago
I'm probably one of the niche guys but I see wider applications for this. I like in dash GPS which I have from an after market unit. I use it frequently. The problem is that to update the maps, you need to purchase a CDROM for a hundred smackers. Once you have it for free on your phone you are not going to do that. So I use the phone. Hooked up, the voice instructions play through the stereo. Direct connection is noticeably superior to blutooth for this purpose. The only problem is the awkwardness of screen control access. I could just mount the phone to the dash but the iPhone screen is a little small. What I want is to connect my phone and then have a nice big touch screen on the dash that functions like the phone. I don't care much about voice control (I don't have Siri) but I suppose it would be nice. I'll look for the videos. It seems improbably to me that they could screw this up but I am often dissappointed

mrpippy 12y ago

> Make decent looking apps for each platform to provide functionality outside of platform supported features. CarPlay and Android Auto both do not allow arbitrary apps (which would be needed to i.e. control the head unit's FM/AM radio or climate controls). I expect this will happen someday, but it's not here yet.

Lagged2Death 12y ago

...non-SmartPhone, non-Apple, non-Google users ... at least in the US this bunch is a tiny minority. "Two-thirds of Americans now have smartphones:" http://www.engadget.com/2014/02/11/two-thirds-of-americans-now-have-smartphones/ http://www.engadget.com/2014/02/11/two-thirds-of-americans-n... One-third is not a tiny minority.

matthewmacleod 12y ago
I suspect it's a tiny minority of the people who are in the market for cars with built-in interactive systems like this, though.
umsm 12y ago
What percentage of new car buyers have smartphones? I have a feeling that most people without smartphones will not be in the market for a new car...
- wcfields 12y ago
  Even if it was 99.99999% I would bet that there's still a huge percentage that has no idea what their phone can do, let alone Bluetooth pairing it. I'd even guess that many people think Bluetooth is the name just for those douchey headsets.

Animats 12y ago

There's also the question of longevity That's one of the big arguments for QNX. The QNX microkernel is very stable. It barely changes from year to year. Versions from 15 years ago run fine. There's no need for OS updates. Cars last a long time. I have a 1985 Ford Bronco in my driveway, and it still has the original EEC IV electronic engine control unit, still working fine. That was designed for a 30 year lifespan, and it made it.

luckydude 12y ago
I agree. I was friends with Dan Hildebrandt, he was one of the (few) guys who worked on the actual microkernel. In my opinion, QNX is the only example of a viable microkernel OS (it's way the heck better than Mach ever was). I love Linux, use it all day, every day, but I can see the wisdom of choosing QNX. It's very stable. The downside is you don't get all the new shiny stuff that runs on Linux without some amount of effort. I suspect Ford doesn't care, Ford has always been more about "it works" than "it's shiny". Says the dude who just bought a 1994 F15 5 speed so his son can learn how to drive a manual :)

a3n 12y ago

How long is any particular flavor of Android going to be supported, vs QNX or some other unix-like RTOS? Yes, you could contract for longer support, but how much current energy will remain with whoever's engineers you contracted to for supporting that older and older version? Also, QNX is an RTOS. Maybe they actually needed an RTOS. How well does Android do as an RTOS?

PinguTS 12y ago

It looks like you have no glue in terms of software development requirements in the automotive industry. Software development for embedded devices is a totally different story than software development in the IT industry. The is no minimal viable product and then deliver experience with some updates. Everything has to work from day 0. Otherwise you will have a big PR disaster. Mercedes had that with its E-Class model W210. Only a few companies have exemptions from that like Tesla. A well established company like Ford, GM, Volkswagen or so who delivers something like 50.000 cars a month over all their different models requires different software quality requirements than a company like Tesla who delivers 20.000 cars a year and is considered as a Startup, which may have some hick-ups. Another issue is, that you have even in a simple car you have at least 15 different control units and the HMI/MMI is only one of these. Throwing hardware at it means increasing prices, which you don't want as a consumer. Also remind you, that development of a car starts somewhere 3 to 5 years before you buy the car. Because the hardware has to be tested very extensively so that everything works in without problems in most worst environment like from Alaska to Mojave desert. Designing hardware with such a broad requirement is not as easy as you may think. That is the reason why automotive rated components are way more expensive than industrial rated components, not even talking about consumer rated components. The problem of what you call "decent hardware" is, that it is only consumer rated from 0°C (32°F) to 80°C (176°F) vs. automotive rated, which is -20°C (-4°F) to 125°C (257°F). Then there are also EMI requirements and much more. The only consumer company that is close to the quality requirements of the automotive industry is Apple. But Apple has a different advantage over the automotive industry, which is they sell millions of their phone within a month vs. a carmaker sells a million of a model may be within 3 to 5 year, depending on the market level (small car vs. premium car). The iPhone within a car is only one ECU, but a car as at least 15 of them (up to 100 in the premium class, like Mercedes S class, Porsche). So the cost structure is completely different. Another thing: Android Auto and CarPlay are no OS at all. Both are just interface definitions. Both are implemented by QNX for example. That said, QNX was one of the first suppliers who teamed up with Apple to provide an integration of the iPod.

superuser2 12y ago

Do not do this. In 15 years, your car will still be around but there is a very good chance that neither CarPlay nor Android Auto will exist anymore. You're going in the right direction - this is slightly less stupid than having something like a Pandora client permanently installed in your car such that it cannot and will not ever be updated once the next model year comes out and Pandora makes breaking API changes or ceases to exist. The only reasonable thing to do is provide long-lived open standard interfaces like the 3.5mm analog connector. Anything else is deliberately anti-consumer planned obsolescence. It doesn't matter if you think "Android is the platform of the future" or something. It probably isn't. Don't tie something with the release-replacement cycle of a car to a disposable device like a smartphone.

twelve40 12y ago
This puts things in perspective, I appreciate that. However, Ford (and others) seem to already tie their cars to one or another type of transient tech. Like the above-mentioned Inrix for maps, or my Pioneer stereo phoning home to some unidentified service to get the traffic. Is Inrix going to be around longer than Google or the 3.5mm jack? I'd rather automakers provided something like a simple, reliable smartphone docking station, with speakers, mics (don't skimp on those like Ford does), and maybe some kind of a display integration, and leave the UI and the smarts to the fast-moving mobile market. But they seem to prefer to force-feed you those crappy, sometimes insanely overpriced systems and web services that still don't seem likely to outlive Google or the car itself.

sandGorgon 12y ago

is this a certification issue ? I know that QNX is certified for several CAN-bus deployments... maybe Android is not certified for those kind of environments. Again not sure if this is part of the vehicle CAN bus anyway.