Short answer to the lede: no, the industry cannot cope. Or rather, it will limp along with bloatware, bugs, and malware exactly the same way we see desktop OSes bloat, or the way we see routers and set-top boxes hacked to become botnets.
In my 40+ years in the industry I've yet to see code get SMALLER. With the exception of Linux kernel 1.0 in the 90's which was a step backwards into smaller, more compact code, code has always bloated.
Damn. I just want a car with as FEW knobs/buttons/levers as necessary. Literally: make it as simple as possible. Like an golf cart! Is anyone else out there with me? I feel like Walter from The Big Lebowski regarding this: has everyone just gone crazy?
Glad to hear I'm not the only one who wants a dumb car. Unfortunately, the idea of electric cars has become bound up with the notion of software-driven cars. I want an electric car with analog controls, no touch screens or over-the-air updates, and minimal software. A car that I feel like I own, rather than one I'm getting a click-through license to use.
You know the dark pattern of presenting a license agreement over and over until it's accepted? I predict one day someone will make the argument in court that they always declined the EULA (perhaps the one in their car) hundreds of times, but one day accidentally brushed "accept" with their finger, and that doesn't constitute legal acceptance, especially since they've demonstrated an effort to decline the EULA, but because of dark patterns, they never can permanently reject the license.
Fun fact - if you buy the car new, you may have the "option" of being presented with these choices before purchase. And you can actually refuse! The dealer will be confused as hell, but they are hell-bent on selling a new car, and will actually void it if possible. I know because I've done it. You don't want the weird spyware (OnStar, Carnet, etcetera)? After agreeing to buy the car, refuse to accept the terms (they legally require your signature for this), and they'll find a way to get around it. If they refuse then go buy a different car, because fuck those guys.
Obviously this will vary by manufacturer and feature, and it's getting harder as time goes on to remove this shit from your vehicle.
I thought I remember seeing some article on dashboard design and how distracting everything is now to driver. I thought I saw a quip in the article about how Honda was going to put out several 2021-2022 models that went back to some kind of really stripped down dashboard?
>If I'm the dark pattern developer I'm not logging when the EULA is accepted or how many times they declined it
This cuts both ways. If the user can show one case of their rejecting the EULA without it being logged, they can then make the claim--correctly or not--that they repeatedly rejected it. If the car refuses to work without the EULA being accepted or rejected, proof of its movement would be sufficient to show repeated rejection.
More pointedly, willfully hiding information like this could backfire massively with the courts or law enforcement.
There is hiding information and not collecting it to begin with.
auto eulaAccepted = readEulaAcceptedFile();
eulaAccepted = promptForEulaAcceptance();
This function reads a file from disk and writes back to it every time. If filesystem write audits aren't enabled there is no way to determine if the most recent write of "true" was the only write.
> If the car refuses to work without the EULA being accepted or rejected, proof of its movement would be sufficient to show repeated rejection.
If the car refuses to work without the EULA being accepted then it wouldn't move. If some functionality was enabled or disabled by accepting the EULA then you would have to show that the functionality was never enabled based on secondhand sources... which would be difficult. Proving something didn't happen is infinitely more difficult than proving something did. Many systems are not designed to handle that level of introspection.
All I'm saying is that the cute legal theory of "I rejected this N times therefore that one time I accepted it is invalid" falls apart for me when I consider that the user would have to prove they never did something, except that one time. Good luck.
> This function reads a file from disk and writes back to it every time. If filesystem write audits aren't enabled there is no way to determine if the most recent write of "true" was the only write.
Yet, that doesn't matter - all the user has to demonstrate in court is that rejecting a EULA causes it to come up again in the future, meaning that sooner or later it is going to get accidentally accepted.
Most people don't seem to understand that the laws are interpreted by human beings that have spent their entire lives studying and applying the law.
Particularly us geeky folk often seem to think that you can find a buffer overflow exploit in the literal wording of the law and the judge will have to let you go. That usually isn't how it works (although the odd case where someone successfully exploits the actual verbiage in the law tends to make headlines and make it seem like that is how it works).
That's not a cute legal hack. Not retaining any information that is not useful and has the potential to be incriminating, unless legally required to, is what every smart company or government agency does. You find the relevant laws, then you write a retention policy.
It's not about information, it's about the user being harassed untill someone clicks "accept" button, which is not compatible with "dealing in good faith"
Do you have proof they clicked it, and not their kid? Did they intend to consent, or click the button by accident/ spilled coffee on the touchscreen?
How can we trust your code at all, if all users say the rejected the contract multiple times, and you claim they accepted it and you are the one that stands to benefit from their acceptance, maybe you are lying? Maybe your system is unreliable?
Is it even legal to use the user's property - i.e. their car - to disturb their peace? If you were to post them a letter, you do it at your expense, but now you are doing it at their expense, they are paying to electricity and data to display those pixels for your benefit.
IANAL, but I'm curious if you are. In my experience, lawyers tend to get a lot done by intimidation. Remember all those ridiculous lawsuits from the RIAA over downloading illegal music? Well, it turns out when people actually took them to court they won every time. But people settled out of fear.
Sure, if the lawyers are actually trying to do things legally first and adhering to the letter of the law, I'm sure they're happy to have their efforts challenged. But I have never, ever, in my life encountered a lawyer who worked this way (and I have lawyers in my family, I've been to court, I've personally hired several, blah blah blah). I am not saying they don't know the law, but they use their greater knowledge of the law to their advantage. Their goal isn't transparency, it's submission.
These were widely covered because most people just settled, and there were high hopes for something putting a damper on the lawsuits... from the page about the second case " It was only the second file-sharing case (after Capitol v. Thomas) to go to verdict in the Recording Industry Association of America's (RIAA) anti-downloading litigation campaign"
So 2 for 2 there were found in favor of the RIAA, for huge damages even after appeal.
As I remember those times, they didn't go after people for simply downloading music, they were going after the people who were actively distributing music on file sharing services.
There's a much harder burden of proof for the RIAA when they sue people over simply downloading songs or albums when the user can easily make the case for fair use. I mean, how many albums, 8-tracks, cassettes and CD's did you buy of the same artist and album? I know I did it thousands of times.
The burden gets a lot easier when you can show the person was actively distributing essentially pirated music to other users - propagating what is by legal definition, illegal activity.
You mean you'll only log that it has been accepted and nothing more?
I guess it becomes their word against... well, nothing, you don't know what they did or didn't do.
For those who care enough to reject EULAs, they should take a few videos of rejecting the license and hopefully that would be enough to shift the onus to the manufacturer, who, as you said, has next to nothing.
I should think that owning a car for more than a few weeks without accepting the EULA could serve as reasonable "proof" that you had no intention of accepting it regardless of what is logged. Especially if the EULA is presented at every startup.
> It was intended that when Newspeak had been adopted once and for all and Oldspeak forgotten, a heretical thought -- that is, a thought diverging from the principles of Ingsoc -- should be literally unthinkable, at least so far as thought is dependent on words. Its vocabulary was so constructed as to give exact and often very subtle expression to every meaning that a Party member could properly wish to express, while excluding all other meanings and also the possibility of arriving at them by indirect methods. This was done ... chiefly by eliminating undesirable words and by stripping such words as remained of unorthodox meanings, and so far as possible of all secondary meanings whatever.
I think consent dialogs are one of the closest things to Newspeak that exist. In any spoken language, new words can be coined to express a desired meaning. In consent dialogs, there is no longer "Accept" and "Decline", instead there is "Accept" and "Ask me later". When presented with an upsell, the choices are not "Yes" and "No", but "Yes, sign me up!" and "No, I don't want to save money."
You probably wouldn't need a log of declines if you had the car for a year and the acceptance was recorded only on day 300. Especially since there will be miles on the car showing that it was driven before day 300...
Activating certain features in Tesla does in fact prompt a EULA (enabling autopilot, full self driving, and ludicrous mode all involve disclaimers which are legally EULA agreements. These pop up after you already bought the car, when you activate them under the settings menu for the first time)
Those are liability releases, not agreements over software licensing. In fact those features actually are sold legally as "parts" on the car you bought, there's no licensing scheme from Tesla yet. Though there is noise being made about offering a monthly license for FSD given that it's at $10k now and lots of people who would want to try it are priced out.
The definition of a EULA agreement is a legal contract that is executed by clicking a button. They often contain liability releases. What software license doesn’t?
> A EULA specifies in detail the rights and restrictions which apply to the use of the software.
> Many EULAs assert extensive liability limitations. Most commonly, an EULA will attempt to hold harmless the software licensor in the event that the software causes damage to the user's computer or data, but some software also proposes limitations on whether the licensor can be held liable for damage that arises through improper use of the software (for example, incorrectly using tax preparation software and incurring penalties as a result).
Mazda has been actively removing touch screens in new models. The screens are still there, but the touch part is replaced with physical controls for cabin features/radio and a puck controller for other stuff.
I know HN loves to hate on touchscreens in cars but I don't really get it. Both cars I own and almost every car I've rented in the last few years (that's quite a few) has had a touchscreen. This is pretty much always how it works: 1) The most used functions have physical buttons and knobs, often on both the steering wheel and the center console 2) Touchscreen is used for uncommonly done things, like adding new Bluetooth connections and adjusting radio settings 3) If the vehicle is moving there are limits on touchscreen use (like the touchscreen will refuse to work after X clicks, or disallow some functions, or both.)
>uncommonly done things, like adding new Bluetooth connections and adjusting radio settings
Adding new Bluetooth connections is something I do more than one would think, because my car doesn't seem to allow Android Auto via a USB cable alone, and periodically it gets into a bad state somehow. The only solution I've found is to forget the pairing and start over. This has happened to me on multiple cars, two Hondas and a BMW, so I'm not sure whether it's universal or not. If for instance, GM cars didn't have this issue, I'd seriously consider one.
Obviously this is a software problem and not a touchscreen issue per se, but I associate it with the touchscreen and the generally poor standards of car tech these days.
Adjusting radio settings is something that I do quite a bit if I am going on a trip, because any given FM station fades out in a fairly short time. I could stream music from my phone, but I have a limited plan due to spending most of my time using wifi these days.
My car does have a reasonably clever and low mental bandwidth way of using the radio, though. There is a "scan" (touchscreen) button and then it changes through the stations reasonably slowly with a "stop" button in the lower left.
I don't understand your third point - it's supposed to be a mitigating factor? It seems like a good reason to "hate on touchscreens" to me, though.
Touchscreens and software bloat are not really central to what I dislike about modern car tech. It's the fact that the user experience has converged on personal computers, where things tend to stop working without any meaningful diagnostic or error and I end up spending 15 minutes resetting and disconnecting/reconnecting things to make it work when I really just want to get home before my food gets cold, etc.
As long as any given computer system is simple enough that the engineers understood it, it seems fine to me in theory.
My 2019 pickup has a backup camera and it's dumb as a post, even though the rest of the truck fairly bristles with sensors. The camera's view appears when I shift into reverse and goes away when I exit reverse.
I think nobody who prefers simplicity objects to that kind of tech. It's overcomplicated interfaces to basic services that we despise, like a volume slider that requires you to look away from the road and that's too easy to mishandle.
So you want a mechanical door lock, manual windows, no car stereo, no power steering, no thermostat, no cruise control, etc...
The problem with that dumb car is that it is missing that very convenient feature. So you want a dumb car but with feature X, because feature X is really great. But the other guy will not care about X and will think it is bloat, but Y is really important, while for someone else, it will be all about Z.
In the end, to satisfy everyone, you will need X+Y+Z, everyone will think it is bloated but you can't remove a single feature without someone complaining... As in, I want things light but don't remove my feature.
Unless it is custom made bloat is almost inevitable.
Push button, engage solenoid, who needs a microcontroller?
> manual windows
Push button, turn on motor, limit switch to turn off the motor when you get to the end. Microcontrollers make pinch detection and calibration nicer though.
> no car stereo
Tons of examples here, a stereo built today would most likely have electronics and/or a microcontroller because it's cheaper than an analog FM radio, but an old one will still work fine.
> no power steering
On my 1978 vehicle. I believe this is hydraulic with a belt powered pump to reduce the user torque on the steering wheel. No electronics, certainly not a microcontroller.
> no cruise control
On my 1978 vehicle. No microcontroller there, just some vacuum linkages. Maybe a bit of electronics to turn things on and off, the switch certainly felt like an electric switch.
None of these things require a microcontroller. Because it's cheaper, almost all of them use a microcontroller on modern cars; thermostats on lower models would probably be the most likely to be a simple system without a microcontroller.
Since when is power steering a "smart" feature? Power locks and windows are mechanical devices, solenoids powered by the car battery. The only computerized feature you listed is cruise control and we've had that for years and years. I don't think that this is an "either / or" proposition where we have either a car running off bloated software or a car limited to 1950's features. As far as I'm concerned, cars from the mid 1990's to mid 2000's are peak.
This is missing the point - Person A wants their next car to have certain features, but person B also wants their next car to have certain features which person A doesn't want. Car companies aren't going to make 200 car variants with different features combinations, they're just going to group all the features people want into new cars and ship it to everyone. To stay competitive they just put out whatever it going to sell and 90%+ of people are fine with the increasing level of touchscreen controls, so they'll keep moving towards that since it also ends up reducing COGS and simplifies assembly of the dashboard, increasing margin. Person A can not buy the car if it doesn’t suit them.
I was going to say something like this, that products (like a lot of things) are a combination of multiple users' wants, but I'd say that the main problem here is that the implementations themselves are becoming more monolithic.
The extreme case (as you mention) is the moment you stick in a touch screen and a fair amount of intelligence. You might as well put a jillion features in there plus you save money on physical controls.
They used to be capable of building cars with numerous option packages in the 1960's and 1970's, it's pretty remarkable. The 1969 Camaro alone had probably a dozen different engines available from 3 families.
Automation drives down costs for standardized, mass produced items, but it may also remove the economies of scale that made it possible for more variety to exist due to decreased demand for those varieties.
> So you want a mechanical door lock, manual windows, no car stereo, no power steering, no thermostat, no cruise control, etc...
I think an extremely small minority of people who want a "dumb" car don't want those features. I feel like you're just making up a strawman here. It's pretty clear (at least to me), that the people who want a dumb car are talking about things integrating with your phone or being displayed on a touch screen. I think anything that existed 20 years ago is not regarded as dumb by basically anybody.
It has been a long time since you could get a truly bespoke car from a major manufacturer in a reasonable price segment.
But, you used to be able to order all sorts of stuff - or not order it. For example, you could save a few hundred bucks by not having a rear bumper, not having a radio, opting for manual windows and transmission, opting for no cruise control, etc...
There were a multitude of options and you could add/delete most anything you wanted.
These days, it's down to packages and colors. Often, you can't even mix packages and trying to get manually controlled windows will usually get you laughed at.
I think you are missing the point. The problem is what do those features mean?
Is the stereo just an AM/FM radio? Does it have a CD player? A CD changer (how many people even buy CDs anymore)? Does it get satellite radio? Does it have AUX input? Does it have Bluetooth? Is there an interface to communicate to the the Bluetooth device (people driving around controlling their stereo from their phone is more dangerous than people doing the same on their car's touchscreen)? Can it stream music without a Bluetooth connection? Does it have Spotify? What about Apple Music? Does it offer handsfree control? And so on.
There simply isn't a universal definition of what a "dumb car" would be and not everyone is going to desire the same set of features.
No I think you're missing the point. I never said there was universal definition. I just said that very few people complaining about smart cars are against features like "mechanical door lock, manual windows, no car stereo, no power steering, no thermostat, no cruise control"...
I can't speak directly for OP, but I think the specific part you are quoting was being facetious. No one would consider power windows as a "smart" feature. But where is the line between a smart power window and a dumb power window? For example, are they just simple windows with an up and down button? Are there options for disabling the window buttons in the back seat for child safety? Are there options for the back seat windows to only go down halfway? Will the windows go down completely with one touch or do you have to hold the button? Can you set the windows to close when you turn off the ignition or lock the car? You might not care about any of those features, but some people will. That is what leads to bloat.
I feel like you're just making up a strawman here. It's pretty clear (at least to me), that the people who want a dumb car are talking about things integrating with your phone or being displayed on a touch screen.
Personally, I really enjoy that with our new van I can listen to a podcast/music through Bluetooth and pause/resume playback through a simple touchscreen without having to fiddle with my phone directly.
> Personally, I really enjoy that with our new van I can listen to a podcast/music through Bluetooth and pause/resume playback through a simple touchscreen without having to fiddle with my phone directly.
I'm not sure what your point is. I never said that no one wants these features. Obviously there are many that do.
Exactly. I don't see why a screen attached to the car is somehow superior to the native screen I've been using for years; just put it somewhere accessible. I've found that the CD player mounted phone holders are insanely convenient in almost every car.
The features you mention aren't any more "smart" than an intermittent wiper is, which was invented long before electronics appeared in cars, much less digital logic.
US luxury cars in 1965 had all the features you mention. They were delightfully dumb and simple to operate. That's what I want now: knobs, sliders, and buttons that move and click when my finger pushes them.
I'll happily take slightly more expensive knobs, sliders and buttons over a touchscreen.
I know exactly where, e.g. my fan speed knob is and I can reach for it confidently. Or I can take my eyes of the road so I'm not blindly groping around the touchscreen to find the increase/decrease button, assuming its even on the right screen and I don't need to go through several menus to find it. Cost doesn't enter my mind in this scenario.
This was literally the pinnacle of dumb cars. Everything could be manual or automatic, but you still got OBDII, warning lights, and of course ABS and air bags. Adding a new DIN head unit, you could have HD radio and Bluetooth completely disconnected from the ECU. I think the automotive industry forgot how to engineer. All the features, all we ever needed in a car, was right there in the late 90s.
Not sure how it is today (haven't bought a new car in many years) but all of these things used to be selectable options a la carte.
It's not particularly difficult for the manufacturer. All the cars had the wiring for all the features since that's the hardest part to do after the factory, but doesn't cost much. Control modules and actuators can be added very late in the assembly line (sometimes even at the dealer prep) so you only get the ones you want to pay for.
You might like the VW e-up!/Skoda Citigo-e. Bare bones electric cars -- even the battery meter is an analog needle! Just has a plastic mount for your smartphone above the center console and a USB port. No giant touchscreens! Real knobs!
Sounds like something that will never come to the USA unfortunately. All new cars are required to have backup cameras as a standard safety feature, so at that point, the car company will ship the whole CarOS anyway.
I've driven rentals in Europe with reversing cameras, in fact my current car had it as an option (I didn't bother). They still had physical controls. Sure the screen is a touch screen too (so when the phone rings you can press green or red on the screen), but the button to select radio, or bluetooth, or whatever is physical, the volume (and off key) is physical, the radio selection is physical (both centre console and on the steering wheel). The dashboard is multiple different guages - there's an LED screen with selectable stats like 'time driving, average fuel consumption, current speed', but there's an nice analog speedo, fuel needle and temperature needle, and several warning lights.
I think the only car I've driven without physical volume controls was a Ford, and that was nearly a decade ago, I get the feeling there's been a bit of a push back, at least in the UK.
Physical controls are vastly more useful - when the control you want available can be planned ahead of time. You get touch feedback when you're operating it, of where it is and what state it's in. You don't have to look. Missing it with your finger is obvious.
Glass controls are optimal for precisely only one scenario, and that is when you don't know ahead of time what will need to be on the screen. That's why smartphones use them.
I own a 2014 Subaru Impreza and my brother in-law a 2017. I've driven both extensively. Sometime during those years, Subaru switched from knobs and levers to a touchscreen and the controls are so much worse!
1. The controls on the 2014 are obvious, easy to find, and my choices are readily apparent. In the 2017 I have to search for them and often guess their meanings. If I'm actively driving, I just give up because it's too distracting.
2. The controls are not as responsive. Sometimes there's a lag. Sometimes they don't respond at all.
3. There are bugs with the digital controls that simply don't exist in the analog versions. As an example in the 2017 the radio turns on every time the car gets started regardless of if it was on when the car was turned off.
Worse, in the 2015 Toyota and 2019 Honda that we now have - if one of us listened to heavy metal on volume 30; and the other one is more of a mellow pop on volume15; they'll have to wait until the car turns on, boots, and timeouts the warning/license messages, before car will accept the volume down / turn radio off input and kill the cacophone.
And yet now the norm.
I'll do you one better - I'm hanging on to my 2004 Subaru WRX for these and similar reasons too :). I go through dealerships every year or two looking for replacement... and keep my WRX with happiness in my heart.
I also hang on to an older car for much the same reasons. Every now and then I’ve had a courtesy car from a garage while mine was in for a service, typically something very recent. I almost invariably dislike them.
You could argue that’s because of familiarity. Obviously I like a lot about my car or I wouldn’t have kept it for this long! But truly, it’s not that I don’t think there’s room for improvement on my old car. There have been plenty of advances since its time that I would welcome if I bought a new vehicle: improved efficiency, safety features, practicalities like better external lighting, and so on.¹
For me, those benefits always seem to be outweighed by the horrible state of controls and displays and “infotainment” systems in new cars. They’re cluttered and intrusive and distracting. I’m not sure which is worse, touchscreens that take your eyes off the road, or cluttering the steering wheel with controls for a phone that it is almost never appropriate to be using while driving anyway in my country. Meanwhile, apparently I still have to take a hand off the wheel for a second or two to change gear or switch on various external safety lights. This should have been some sort of meme punchline, not real life.
Between poorly designed controls, the security and privacy problems that seem to be rampant in modern cars, and the sense that EVs might be the future but there are still some big unanswered questions, I expect it will be a few more years before I change to a new vehicle unless some practical consideration forces the issue first.
¹ I’m still waiting for the windscreen that automatically enhances the driver’s full view in low light or poor weather conditions and subtly highlights the required driving line through junctions as directed by the navigation system, but I reckon we’ll have that too before the self-driving flying cars are here.
The RAV4 we had for 4 years now - I'm thoroughly familiar with the UI, and parts of it are still completely non-sensible.
Like yourself, before COVID I traveled and rented cars frequently; there are some UI choices that are inherently poor or against my priorities/workflows and they'll never be right. I could list them, but now we'd be into rant territory... :-/
I have a similar problem with my 2017 Pacifica that replaced my 2005 Voyager. Using the touch screen to control the heat and air-conditioning is extremely slow in the Pacifica, and you can't do it without looking at the screen which can't be used with gloves.
There are physical buttons and knobs for a few of the controls, but it seems that they're just talking to the same software as the touch-screen, so they're just as slow to respond. Adjusting the heat without looking at the touch-screen is pretty much impossible.
IMO the touch screen should not be used to control any aspect of the car's operation. It should only be for phone, navigation, backup camera, and entertainment.
It might depend on the trim of the Pacifica? But for the hvac controls that have buttons (fan speed, driver/passenger temps, mode) the physical controls seem responsive enough and I don't use the touchscreen ones, except maybe sometimes for defrost if I can't find the button; cycling through the modes can be painful. You can't get the different zones back in sync without the screen though, which is annoying.
And going from lo to hi takes forever, too. At least the fan speed knob seems good enough. The touch screen in general is decidedly not great; it's better than my C-max (sync2) in many ways, but I really like how the sync2 was designed --- they clearly considered how to make it useful, but then implemented it in the slowest environment possible.
I also have a 2017 Pacifica, and for the most part I don't mind the controls (the physical buttons are responsive enough that it doesn't bother me), but there are definitely a few functions (heated/cooled seats, mainly) that I wish I didn't have to dig through menus to find. Such a contrast from my 2009 Civic (albeit no heated/cooled seats on that vehicle).
The console is awful. Truly awful. Push the volume button within a few seconds of turning the car on to turn the radio off? Doesn't even register. Even if you give it a minute to warm up, it is half a second of lag on a physical switch. The touch screen has a half second of lag and requires several presses to do something as simple as switch from the radio to bluetooth.
OT but since there seems to be multiple Subaru owners — if you are experiencing an issue with the clock on MFD(the small display on top of the dashboard) being too fast, probably depends on models but the causes are,
1) that there’s no CAN bus messages in Subaru cars that offer GPS time, and
2) that at least older models of MFD counts time by dividing 125kHz CAN bus crystal, where a sane choice would be to use 32.768kHz one.
It’s not just your car, the issue is in design. To hypothetically fix it a firmware hack would need to be built. I learned this when a friend of mine told his is always way too fast and often makes him upset for a moment that he might be late to work — don’t know he meant it justify flooring it but sounded like he was genuinely annoyed.
I have a 2016 Outback and love nearly everything about the car, except the stupid console. Just horrible in the ways you describe. Laggy, unintuitive, and just irritating to use. I only ever use the actual buttons on the steering wheel. Not only is the console touch screen bad, but the actual buttons for the HVAC system are weird and unintuitive. 4 years of owning it and I still push the wrong buttons.
It's still an issue in the 2018 Impreza. I've gotten used to it now, but it's still frustrating. Especially if I ended the drive with a phone call. On phone calls I have to really crank up the volume (nearly max) whereas for music and other things I have it very low (10-15? The numbers mean nothing to me, not loud). So when I turn it on after a call I get blasted by NPR for about 5-10 seconds before the audio volume dial actually responds (sometimes it takes the early attempted dialing down but delayed, other times I have to try and dial it down again).
Other than audio controls, though, everything else is responsive, I don't notice any lag. It really seems to be an issue with their stereo system. Either it's not fully booted (and can't respond to controls yet) or there's some mediating system that transmits the controls which isn't booted up as quickly.
I have the same problems with my 2019 Impreza. It responds sooner if I'm not in reverse, but I am usually in reverse first because I need to back out of my garage. It usually doesn't respond to the volume knob messages until a second after shifting into forward. I say "messages" because the turns of the knob appear to be queued up somewhere, but it doesn't have as much time to process those while it's displaying the rear-view camera on the screen.
I have a problem like this in my 2018 Dodge! The bluetooth takes a wildly variable amount of time to connect when starting the car, and the bluetooth doesn't have an input gain option, like aux input does.
This means I need to turn the volume up a ways to hear anything over bluetooth. This also means that between starting the car and bluetooth connecting I'm getting blasted by some random radio station at increased volume.
Huh. My 2020 Impreza has lots of knobs and levers.
About the only time I interact with the touch screen is to control the apps I use (mostly maps and podcasts). Even the podcasts app rarely requires me to touch the touch screen; there are volume and back/forward switches right on the steering wheel. (And also on the console below the touch screen.)
Maybe that's switching back after 2017, or perhaps you use more of the controls than I do.
Apple had those keys in the same place for ages, and they're all set to the the action rather than function by default. Volume up and down is useful, as is escape and the power button. Key brightness is handy, play/pause I used to use a lot. So looking forward to finally getting a new laptop when the M1 Macbook Pro comes out.
Kinda not really - yes you can kinda see it - but you need to look at it most of the time to hit a button correctly. Which you didn’t always/usually have to do with physical buttons. It also switches any time there is a context switch, which depending on what is going on can be insanely distracting (especially when it is using it to display autocomplete, autocorrect suggestions as you type fast)
I understand why buyers like digital novelties in cars: they're flashy and sexy, and they make your 4 year old car look old by comparison.
I also understand why car makers like digital flash: your 4 year old car doesn't support your latest iPhone, network, or peripherals, so you're motivated to buy a new one every few years.
(And of course, old car tech distracts your driving LESS, however that fits into the picture.)
I think we will never see cars with modular digital tech that can be updated. A car with replaceable digital hardware and software won't rapidly go out-of-date the way current cars do. They would cost far less to update than replace. So nobody wants it... except perhaps grownups.
I recently bought a used 2016 Spark EV (having a baby and needed something other than my motorbike). It does have a touchscreen, but mostly for extraneous information and radio functions. Everything else has dedicated knobs and buttons. (The one dumb thing is that the fan speed updates on the screen, rather than just having ticks above the knob.)
It is a California "compliance car", which means it was just a modified petrol Spark, so it didn't have product managers trying to jam in unnecessary touch-based interfaces.
I just hope that all companies building electric cars don't move in the all-touch direction of Tesla.
never mind that when I'm taking a photo outdoors on a bright sunny day, there's no way in fuck I'm going to be able read your LCD screen.
Never mind us older folks, who often have vision problems with up-close viewing, which is a solved problem when you can spin a shutter or aperture wheel by touch. But not when you need no-glasses to view the objective, but reading glasses to view the fine print on the screen.
I was using those dials all the time as a 16 year old learning to use a camera. It's really easy - especially when you can set one of them to auto and just control the one you care about - plus so quick when they're on a dial.
But I can believe no one wants to do it now, it's amazing how much a bit of tactile feedback can make a task 100X easier, and something you can do instinctively.
Tactile feedback and consistent UIs are great and a common feature (if not a defining one). I've got a Pentax dSLR -- it's got an excellent UI (better than canon's, imo) and their manual tweak focus adjustment & focus hold system does just work very well. It also has an excellent set of manual controls. A pity that nobody else has heard of them...
Which DSLR no longer has an aperture and shutter speed knob? I know the MFT mirrorless lost most of the controls, but did the latest Canon/Nikon really do this? My DSLR is eons old, so this is just baffling to me.
My Nikon D850 (DSLR) and Z6 (mirrorless) both have separate physical dials for shutter speed and aperture. The D850 has buttons that allow the physical command dials to be used to control ISO, white balance, bracketing intervals and load of other things. Important settings are shown in the viewfinder and top panel and I only really use the rear monitor to check the histogram for over exposure. I almost never need to use the menu system / touchscreen to alter a shooting setting.
I have an older Canon 5Dmkii, and it has a dial for shutter speed near the shutter release and also has a dial ring for aperture control on the back. I'm pretty sure the mkiii and mkiv do as well. I know the 1D I used in the past also had the same configuration of dials.
As with anything camera related, where to buy could be B&H, Adorama, Sammy's, or your local camera shop (if they still exist in your area).
I actually used to work as a test driver for FCA, and we were specifically told to report on knobs/dials not working or wiggling too much. Our training had us touch literally everything in the car that moved. Albeit, I'm not sure how much of that feedback was actually acted upon, based on the quality of some of those cars.
It annoys me about capitalism. There's a massive incentive for manufacturers to cut corners to save $100 on a $30k car, as they sell 10,000 cars and use the $1m to pay themselves a bonus about how great they are.
Almost everyone will want to pay $30,100 for the better product though, but the market can't differentiate on that.
I drove an Electra Meccanica Solo in Victoria BC a few years ago, and it was exactly this. The founder of the company wanted to design the air cooled Porsche of EVs. All analog, minimal electronics, incredibly nimble. One of my favourite aspects of it was the sound it made under acceleration, think: Star Wars Landspeeder.
(You can faintly hear it over annoying background music in this video .)
It felt better to drive than a Tesla, even though it was orders of magnitude slower. Unfortunately, last I heard they were trying to tone down the Landspeeder-esque “cabin noise”. I’m not sure if the current executive team at Electra Meccanica even realizes what they have.
One way to get that is to do an EV conversion. Unfortunately, it tends to be expensive and time consuming, but in the end you get something that's just a car and not somebody's notion of an ideal consumer electronics experience.
That's not to say there'll be no software, or even that it won't be proprietary. But at least it's a lot nicer to buy a motor controller, a charger, a battery management system and whatever else you need from companies that know the parts have to be easy to configure and interoperate with other components from other companies, because if they aren't people won't buy them. Modularity is a wonderful thing.
(I just ordered the charger and BMS for a project I'm working on yesterday. I ended up opting for more expensive components because they're configurable by the end user, whereas the popular cheap option you have to send it to someone to reprogram if you change your battery pack configuration.)
I really miss "real buttons", meaning not just something tactile. Things like a power button that actually opens a circuit. Or a volume dial that doesn't lag because it's actually a potentiometer and not a rotary encoder. Too late for all that, I suppose.
Ha, as much as I want simplicity, I don't miss switches and potentiometers with burned contacts that cause intermittent failure. I'd be happy with reliable optical encoders and properly embedded digital systems. The kind that can hum along for decades without maintenance...
I think this way of thinking should lead people to take another look at bikes, perhaps power-assisted electric bikes. Cheap, simple and efficient in comparison to cars. Not to mention all the benefits of cities made for bikes vs cities made for cars.
I want a smart car. I just don't want the smarts determined by the manufacturer. I really want a car that can more or less run arbitrary software/features I want. I understand why companies do not want to provide that, because people can go the very wrong direction with it. But I'd like both my house and my car to be smart, and neither beholden to Apple or Google to do it.
Same here. The automakers are trying to pull a fast one by conflating battery power with tons of licensed software with an attack on independent mechanics and serviceability. But I suspect that there is a huge market for low end EVs that can be easily serviced and don't have a lot of software.
There are compliance cars that are just electric versions of the normal ICE cars that preceded them. That's probably as close as you can get. I currently dive a Bolt and it's a bog standard car, just electric. Yes, it has a touchscreen, but that's for infotainment.
I have a more optimist view about the industry being able to cope just because we have two more industries that have gone the same transitions, aeronautical and aerospace.
Back in the day airplanes where just knobs and levers, and we didn't have the reliability that safety that we have today.
With aerospace, I mean, software engineering as a discipline started with aerospace!
If we write "starup code" (i.e. CRUD web app) for cars we'd still be in huge trouble, but if the automotive industry can adopt the redundant systems, enforce VERY high levels of software testing, and other practices for those other industries I think we could see some interesting things coming from the industry
I'm sure you didn't mean it quite so bluntly, but characterizing the history of aviation safety as a triumph of computer control systems over analog ones misses the mark by a pretty wide margin--even if we're just talking about the advances that only computers have provided. Probably the only way a modern car is technologically less sophisticated than a passenger aircraft is its inability to substantially steer itself (and this problem is orders of magnitude more difficult for cars than planes).
A lot of the systems we take for granted in our cars (ECU, ABS/stability control, adaptive cruise control, steer-by-wire, OBD) were pioneered in the aviation industry, and both industries' safety records have been massively improved by the ability to do digital design/analysis (CAD, FEA, CFD, etc). Then once you start talking about advances in computer-aided manufacturing and QC/QA processes, training, failure analysis, human-machine interaction...
One of my perennial frustrations with current tech is the idea that putting a computer in the control loop necessarily makes things safer.
> One of my perennial frustrations with current tech is the idea that putting a computer in the control loop necessarily makes things safer.
I completely agree with you! A simple electronic component is a lot more fragile that a mechanical counterpart that is unfazed by ESD, vibrations or whatever other things that can kill a electronic component, or a circuit board for that matter.
My point is more along the lines that a computer in a control loop makes things different, not necessarily safer. But with the flexibility that a computer brings to the mix, if used properly a computer can add some safety features that would be hard to implement with only analog/mechanical parts.
It seems to me that we have reaching a ceiling with what we can do with mechanical systems, although I do believe that we often get lazy and opt for software convenience instead of using mechanical reliability where it would be beneficial.
So all in all, computer control systems are not safer just in themselves, but they can be, if not in reliability, at least in monitoring health and providing warnings before things are critical (i.e. a temperature reading instead of waiting to see smoke coming of the hood of a car)
> A simple electronic component is a lot more fragile that a mechanical counterpart that is unfazed by ESD, vibrations or whatever other things that can kill a electronic component, or a circuit board for that matter.
> I have a more optimist view about the industry being able to cope
TBH, I have similarly optimistic views, but for completely different reasons. The truth is, the worldview has changed completely in the past 100 years.
In the old days, if you failed at operating a saw, you hurt yourself badly. Now we slowly are starting to expect sawstop and other solutions to reduce injury.
The same thing is happening in cars. We no longer expect perfection of the human as we augment them in various ways to both reduce the frequency and severity of collision. Its the early days yet, and its very much the "startup code" mindset with cars having way more bugs than its ever but also producing safer outcomes.
Car software is getting worse, but we're better off for it.
Car guy here. Yeah, I’m with you. My favorite car was a ‘96 Tercel with mechanical steering and a 4-speed manual. It even had manual roll-up windows. It had less than 100 HP, but it was simple as hell to operate, very fun to drive, cheap to maintain, and extremely reliable up until I sold it with nearly 400k miles on the odo. I put an aftermarket stereo and speakers in it and I was set. Only reason I sold it was because my wife hated it and a friend needed a cheap reliable car for his idiot son who proceeded to neglect and destroy it quickly after taking possession.
Modern cars are absolutely terrible in the UX department, but they are a hell of a lot safer, so there’s that.
Those mid 90s Japanese sedans were awesome. Relatively compact overall, zippy little 4 cylinders engines and manual transmissions. I had an Accord from that era, and loved driving my dads nerdy-as-hell Nissan Sentra. That car was shockingly fun to drive.
> "Modern cars are absolutely terrible in the UX department, but they are a hell of a lot safer, so there’s that."
tangentially, "safety" is highly cargo-culted. things that seem so obviously safer are taken without question as better, but in many cases, such features really only provide a false sense of security along with substantive unintended consequences.
most safety features in cars (e.g., lane-keeping) allow people to be less skilled and less attentive at driving, rather than lowering crash/injury/death rates. the better solution is to make people better and more attentive at driving through more rigorous training/testing, more thoughtful design, and importantly, culture, rather than just technology for its own sake.
Modern cars have more design features that improve crash survivability (better airbags, better crumple zones, tested with more realistic crash tests, etc). Those seem like a pretty unalloyed improvement to me.
I'm not saying you're wrong about the things you mentioned, but cars really have got safer, in important ways.
yes, airbags and crumple zones do improve safety, but even those are not without negative consequences, like bigger, heavier vehicles (which is more dangerous to others) and higher sense of psychological safety leading to being less considerate, less attentive, and more reckless.
that's not to argue that those tradeoffs aren't net positive, but that they're still tradeoffs to be considered, rather than short-circuiting to "of course it's better!".
if you’re prone to being overwhelmed by stress while driving a machine that can potentially kill you, that’s a sign to get more training, not to mollify oneself with an illusion of safety. ignorance is not bliss in this case.
Have you ever seen how stressed out drivers behave?
We've all seen that person in a Honda fit who's so terrified of being on the freeway during rush hour they're driving like an overloaded scrap hauler and generally causing a problem in whatever lane they're in.
Unless you crash into something going highway speed a car from 1990 or 1980 is fine as long as you're wearing a seatbelt.
The states look good because safety tech that only matters at the extreme end has improved so the drunks and distracted teenager who would have dies had they gone off a cliff in a 20-30yo car are now surviving.
Basically the medium to low speed impacts that people not behaving particularly poorly get into have always been highly survivable.
This is too dismissive of the many ways people get into accidents. I got t-boned by someone who blew past a stop sign on a residential road. Side curtain airbags prevented me from hitting my head on my side window.
Those are two different stats - parent was referring to if those features lower death rates. You are looking at total death rates regardless of features.
It’s also possible better road maintenance, or airbags, or better crumple zones (but not lane help) are driving it down. It’s possible for lane help to be driving it up, just not as much as say better crumple zones, and it will still be trending down.
> things that seem so obviously safer are taken without question as better, but in many cases, such features really only provide a false sense of security along with substantive unintended consequences
This is not true at all. There is definitely testing of cars by groups like the NHTSA and Insurance Institute for Highway Safety. For example, they've found crash rates were 14% lower on cars with blind-spot monitoring. Modern cars are far safer than older cars.
The nice thing about those cars is that you can actually replace things like suspension and brake systems with modern equivalents relatively easily. It’s not like swapping in an F-150 brake system into a Corolla.
As far as rust, I’m sure you can find some rust free cars in AZ or CA.
God I hate that video so much. It is right up there with the Top Gear Hilux episode when it comes to entertainment being peddled as scientific experiment by people who find it convenient to the point they are trying to make to do so (not saying you did that, just that tons of people do). It was a promotional video designed to highlight progress. It does that just fine. But every two bit tom dick and harry seems to think it's representative of a crash between any old car and any new car when the vehicle in the video doesn't even have seat-belts (!!!!) which are the single biggest improvement to crash safety/survivability ever.
There's a reason nobody is going to do a crash test of a 1980s Ford cop car into a 2010s Ford cop car. It would be a massive nothingburger.
Frontal airbags do little for seatbelted passengers at crash test speeds (texting into the back of a semi at 70mph is a different story). In fact in many cases they are worse than no airbag at low/medium speeds because you injure your face hitting airbag or are injured in its deployment in situations where you would have otherwise hit nothing. They are intended as damage control for unbelted passengers (which is where do you think the first S in SRS comes from).
Crumple zones are highly overrated. Imagine an old cartoon where someone gets thrown off a building and happens to land on a guy carrying mattresses. That's basically how crumple zones work. Too high of a jump or too soft of a mattress and you blow right through with negligible deceleration. Too hard of a mattress and the mattress decelerates you basically as hard a the sidewalk. Crumple zones provide time for the airbag to deploy, yes but they only really mitigate the forces on the passenger cabin (and its occupants) at a narrow speed range (they shoot for the speed of the relevant crash tests).
For people who were doing things right and wearing their seatbelt the safety of crashing a car changed negligibly from the advent of the 3pt belt to the time of the side curtain airbag. (I'm ignoring rollover safety here, that's been a basically linear improvement since the 1960s but it's not technically interesting because it's a simple case of throwing more and more steel at the problem)
Side curtain airbags have been a major improvement because there's not much restraint nor space for deceleration in that direction. They do a great job keeping people's heads from hitting stuff.
AEB is another good modern one. Knocking 5-20mph off the speed of a collision helps a lot no matter how you cut it.
All that said, everyone who worships safety tech uncritically as though they're all big improvements you can stack on top of each other and get the sum of them all needs to take a long walk off a short pier. Professing your love for airbags and crumple zones might get you fake internet points on Reddit or HN but that's just not how it works and those people pollute and/or prevent legitimate discussion.
I get what you are saying but look at what happened to the 60s car here. Seatbelts don’t do anything for you if the other car is suddenly parked in your pelvis. None of the hard parts seem to reach at least the torso of the dummy in the modern car while in the 60s car the dummy ends up being pushed virtually into the back seat. Sure the seatbelt would have held it in the seat but it that does nothing as the bent door and the grill of the modern car crushed it’s lungs and liver areas. The airbag here wouldn’t have saved the dummy either of course. The thing that would have is a better crumple zone.
Most safety features are nothing like lane assist. Lane assist is not a representative example of what most safety features in a care look like. Most safety features are crumple hoods and pillars, safety glass, airbags and seatbelts.
Additionally, you think it is a more feasible solution to change culture than to make the safe operation of a car less reliable on the human element? How exactly do you "make people better and more attentive?" Training? Please elaborate on what that would look like.
Cars are designed a lot more thoughtfully than they used to be. An example not related to safety is cupholders. Remember when two flat rings on the dash were considered acceptable for holding a cup? Safety wise I'd say designing contingencies against inattentive drivers is thoughtful design.
The "spewing BS is order to show the world how much you care about safety" part of the safety cargo cult bothers me a hell of a lot more than safety features enabling less attentive behavior.
If I had a nickle for every social media user who's screeching about how airbags are all that's good in the world but who doesn't understand the inherent tradeoffs to a bag of dense air occupying the space in front of the driver or how a crumple zone is basically a single use spring I would use the money to hire people to beat the first group of people over the head with highshcool physics textbooks.
I wish we'd regress to '50s attitudes about safety simply so that the people who care about safety no longer have to be held back by the people who crap all over everything by wanting to be seen caring about safety.
yes, it's frenzy, a typically pro-social behavior (agreeableness) that's misapplied due to fear and uncertainty, so that we get counterintuitive consequences from overly correlated belief/behavior.
it's the same impulse that got us mis-masking and lockdowns, when the real danger was inter-/intra-familial contact (because that contact is prolonged and our guards are down, unlike with "stranger danger").
happens all too often, and it's frustrating to watch otherwise perceptive and reasonable people forsake those facilities for primal herd behavior.
I have a 97 Miata and a 95 4Runner. Both high mileage, both amazing cars mechanically. They both have power steering and power windows but everything works on both cars: every knob, button, window, still works. I've had several newer cars that have degraded much more quickly than these two. Both cars are much better to drive IMO than newer cars, with some caveats. You can't drive like a dummy and expect these older cars to kick in computerized traction control systems to save you from yourself. I personally like a car that lets me be the driver.
>My favorite car was a ‘96 Tercel with mechanical steering and a 4-speed manual. It even had manual roll-up windows.
I had a 1993 Tercel, same gearbox, manual windows, vinyl interior. It had a leaking head gasket when I bought it, and I drove it for 150,000 km without putting a dime into it outside of oil changes and tires, filling the coolant as needed. Simple cars are basically indestructible.
Counterpoint, I'm not a car guy and I think too many people own them. But when I'm in a car my priorities are safety, mileage, seat comfort, climate control, sound quality, and smartphone support. I don't really care how much it costs to maintain, and I haven't changed my own oil in over a decade.
Modern cars are fine in the UX for what I do. I'd rather not own a car than drive manual, and even vehicles from 10 years ago aren't competitive in creature comforts or gasoline consumption.
I find it more stressful to drive automatic. You need to be more careful with the gas pedal to get just the right amount of acceleration, since too much will make the car shift down and rev up. Also there is the risk of uncontrolled acceleration by mixing up the gas and brake pedal and not having the clutch as a double safe.
If you want the simplest car possible, you can look at models designed to sell in volume worldwide. I'm talking the Hyundai Venue crossover, Hyundai Accent subcompact sedan, Honda Fit/Jazz hatchback (recently discontinued for the US), Ford EcoSport crossover, etc. They are designed to be serviced in poor conditions. They will tolerate removal of electronics like the infotainment because some countries' base models have simpler configurations. In the case of the Honda Fit, the climate control dials physically move the ducts! And repair documentation and parts will be plentiful because there's a large market for parts suppliers to compete. Also look at body-on-frame fleet vehicles like a base Ford Ranger or F-150, but be ready to forgo stuff like cruise control.
There are EV equivalents too, like the Chevy Bolt and Hyundai Kona EV that offer range in the 200s of miles with make driver assists as an optional upgrade. Even with complicated powertrain electronics, EVs are still more reliable than ICE cars because any iffy software is made up for by lack of mechanical parts. The repair procedure is the same as mechnical parts - just swap the faulty part out for a working one, and the "upstream" supplier will probably take the broken module to reflash software or frankenstein together half-working PCBs to make another refurbished module to sell.
If you're worried about remote compromise, it's pretty easy to avoid IMO. Open the dashboard and yank the cellular antenna, and never pair the infotainment to Bluetooth or Wi-Fi (or just yank the 2.4 GHz antenna). Those are basically the only avenues for wireless attacks into a vehicle unless you count the TPMS and key fob radios, which seem too simple and low-bandwidth to offer an attack surface. And if an attacker can access your car's physical ports, they could already attack you in other ways like by weakening the brake lines. Other new electronics, like MOSFETS instead of relays in the BCM, have actually made the car more reliable so they should be fine. Other newly standard features like blind spot monitoring are (1) solid state, so they won't fail often and (2) tolerate failure or complete removal.
They're not objectively better. They're more suited to what they want: a device with fewer features but more direct control.
Lots of people want a car with more features, and don't wish to control it directly. They would rather have a solid-state device that can't be repaired, but doesn't need to be. They don't want to upgrade it themselves; ideally, they don't want it to need upgrades until they buy another one.
But you've got it exactly right: many consumers want the car industry to go iPhone, for the same reason they want iPhones. That's neither objectively worse nor objectively better. The only objective thing is that a ton of consumers want it, because it suits their needs. And part of that is achieved by avoiding development by the kinds of techies who think that their preferences are objectively better.
Some people want their cars to be like iPhones, but is that really what's behind this?
I mean, take TVs. Pretty much every TV has "smart TV" features embedded, to the point where it's hard to find one that doesn't. I suspect this is because the manufacturer gets some kind of kickback from streaming services to include their app and to show ads and to gather data about what people are watching and in the end it's actually cheaper to include an embedded processor than it is to leave it out and lose their kickbacks.
Similarly, those "smart car" features can be monetized by collecting valuable data and introducing subtle suggestions on where to go. It's hard to imagine a car company seeing the kind of revenue that Facebook and Google pull in and not wanting to get in on that. And they can even justify it as good for their customers by saying "we need these computers to collect training data so the self-driving features we deploy in the future will be better and safer."
I can imagine a future where "dumb car" features are actually luxury features. Only high-end cars lack the smart car features because it's cheaper to include them than leave them out. Maybe that already describes the present.
I recently purchased a Kia Telluride. Perhaps I'm now biased because I love the car, but I think it has an excellent combination of tech and usability.
1. It still has buttons and knobs for the things that should be buttons and knobs (e.g. climate control, volume, etc).
2. The heads up display is the killer feature that should be standard on all cars (as is only available on the top speced Telluride). It displays speed, speed limit, blind spot monitoring, lane departure, automated steering, navigation, etc. Unfortunately I don't think Android Auto or Apple Carplay's navigation can be displayed on the HUD - only the OEM Kia nav.
3. The "Smart Cruise" aka Highway Drive Assist aka Adaptive Cruise Control is essentially self-driving minus lane changes. It's engaged with a single button on the wheel and presented in the HUD. It takes corners smoother than I'm able to. I often feel it's turning too early and fight the auto driving but in almost all cases it's correct and my inputs are delayed.
Kia's mobile app for remote start, climate control, valet mode, etc. is pretty terrible though.
2021 Sonata (Limited) here and I echo your sentiment. I spent a lot of time and realized that only Hyundai and Kia provided a good mix of the things I wanted in a smart way.
Mazda's anti-touch stance makes it a non-starter for Android Auto/CarPlay
Toyota, Subaru and Honda have abysmal infotainment interfaces and didn't seem to invest in me enjoying the interior of a car.
Nissan CVTs have questionable reliability . Etc.
The lane keep assist on other average vehicles is so laughably had compared to what Hyundai and Kia have pulled off. These systems actually keep you centered instead of bouncing between lanes (hello Mazda).
Blind spot cameras, birdseye cameras, HUD, little touches like the car slowing down automatically for you on some mapped curves when using cruise control etc. (Also I've used that stupid smart park thing waayy more than expected. I love it)
I'm still disappointed that car reviewers don't focus on such usability issues and instead rag on about performance of the engines or "its a toyonda so it is reliable" and ignore such small but practical tweaks.
Mazda's anti-touch stance makes it a non-starter for Android Auto/CarPlay
FWIW my daughter has absolutely no problem using CarPlay on her CX-30. She quickly breezes thru the menus like it's a twitch shooter game. She can probably navigate faster than if it were a touchscreen, because the control pad is low. Her hand isn't extended way out on the dash.
I purchased a new C8 to add to the collection until my granddaughter is old enough to drive. The HUD is very, very good. I'm not sure how they managed on such a sharp windshield, but it's crisp and clear even in the brightest of conditions.
With them making so many things mandatory these days, I'd not be surprised to see them making HUD mandatory. It absolutely is easier and faster to read then just looking down. It may only be measured in milliseconds, but it's definitely faster.
Specifically, the post I was replying to was lamenting a future where EVs all try to be like Tesla, putting nearly 100% of the instrumentation and controls into a single large touchscreen.
This is unique to Tesla, AFAIK. It worked out very well, it is a brilliant cost-saving measure that can also be marketed as 'elegant'. Have to hand it to Musk on that one. The R&D required to build a normal car interior is expensive, so skipping most of it was a phenomenal move.
So anyway, my point is that there are lots of non-Tesla EV choices, and by-and-large almost all of them are just normal cars with normal dashboards and normal controls, just with electric power instead of gasoline. A few manufacturers seem to be trying to ape Tesla's design on their EVs, but personally I expect a reversion to the mean. Too many people could care less about driving a touchscreen.
Musk coming from software they seem to actually adopted modern software engineering practices that other manufacturers going to struggle with for some time. Being vertically integrated helps heaps too.
Tesla cars have a distinct look and a "prestige" whereas something like Ford's new F-150 will look nearly identical to its gas guzzling brother. There's an ego carried behind Telsa's branding and marketing that other car manufacturers don't feed into even in their own EV offerings.
I think what they were getting at is: there's a minimal difference in terms of "look and feel" between non-tesla EVs and the gas counterparts designed by the same manufacturer.
Tesla very much intends to redefine the driver experience as we know it. In the model 3, the entire control system in the center of the dashboard has been replaced by a giant singular 15" carputer touch screen.
If you don't believe me, you can make the comparison for yourself. Just do an image search for the interior on the Model 3 and compare it to that of a Nissan Leaf or Chevy Bolt.
>soon enough after that we will be able to build electric cars from kits
An internal combustion engine is a complicated, engineering marvel. But a complete engine, as a unit, isn't difficult to remove/insert. It's big, heavy and awkward, sure (so are electric motors and batteries), but an experienced person can do an engine swap in a couple of hours. And yet there aren't many people building ICE cars from kits. Why not?
Because the drivetrain is only one part of what makes a good car good.
And yet there aren't many people building ICE cars from kits. Why not?
Quite frankly, speaking as someone who is a bit of an automotive enthusiast, you'll probably find even less EVs, because EVs are sterile and boring. They can definitely be fast, but I suspect the average auto enthusiast is not only interested in speed. The sounds and smells of an ICE are far more appealing to the type of people who tend to build custom cars.
>And yet there aren't many people building ICE cars from kits. Why not?
There is basically no market for kit cars because if people want a particular kind of vehicle it probably already exists. If they want a particular kind of powertrain it probably already exists. So it's just a case of getting them and combining them. And there are kits that make the most popular combinations a bolt in deal.
A few years ago I found a source selling EV conversion kits for vintage vehicles. It’s definitely a possibility to DIY an electric car.
I could imagine it being like DIY synthesizer kits. Do it yourself if you have the skills and time to put it all together, or spend a bit more for it to come pre assembled, or “some assembly required”.
I would be curious about the safety regulations behind this.
I wish EV tax credits applied to conversions and not just new vehicles. It would make conversions a lot more economical, and maybe we'd start seeing low-cost mass-produced conversion kits designed for popular vehicle models.
May I ask why you think that? (Sincere question, not meant to be snarky). Even assuming electric car assembly is simpler than an ICE powered car, there are still a lot of large, heavy, and expensive components. And that's setting aside all the regulations around it being a "street legal" vehicle, which can place odd constraints and vary from location to location.
I could see there being ways to build your own, just as you could build your own house or laptop if you acquire the right parts. What makes you think that people will broadly want and use kits? (or am I misinterpreting what you're saying?)
The overall mechanics of an electric car is (or can be made, if it's not a conversion) so much simpler; the battery is, in fact, modular, each piece weighing about the same as or less than a regular car battery. I can easily imagine a market for consumer EVs assembled not by huge plants but rather by small local shops as well as enthusiasts. The age of the large brands dominating is coming to an end...
Some of the extra infotainment stuff (read: media content) is subscription only. Some functional extras like full self-driving or faster acceleration are one-off purchases. The rest of the software -- all the things you'd expect the car to do -- comes with the vehicle and gets regular OTA updates.
I think with Tesla the only subscription is "premium connectivity", which is basically a cell data plan. The car works fine without it. Even the nav will still use traffic data for routing, it just won't visualize it for you.
In addition to the comments about premium connectivity, I don't see how the acceleration unlock is anything but a positive. Car companies have been charging for options for forever, except before you had to make the choice at purchase time. Now the upgrade can happen whenever.
As a Tesla owner, I’ve slowly come to realize that most complaints about Teslas online don’t reflect the reality of driving one. I was skeptical of all the things people are complaining about here, but I decided to take advantage of the seven day return policy and drive a Mode 3 for a week: I discovered that nearly every control I care about is controlled by one of two sticks or two little thumb wheels and I almost never have to interact with the screen while driving. Similarly, as others note here, the subscription is for the cell connectivity and media packages, not for the “car’s firmware”. Finally, as a software developer, I really appreciate that Tesla has made updating my car’s firmware trivial and they actually ship firmware updates to older cars: there’s much less difference between a 2018 Tesla and a 2021 Tesla than between similarly spaced cars from other manufacturers.
Two months ago I sold a dead-simple Toyota truck I'd driven for 14 years and bought a new, loaded Honda. The difference in the complexity is of course astonishing. The yota didn't even have usb. You could climb into the engine compartment and sit there. The Honda has pretty much everything a modern car can have, and you can't even see empty space in the engine compartment.
I _could_ get on your "just give me simple" bandwagon. A part of me was proud of the low-tech, rusty durability of the Toyota, I admit it. Still, I'll also admit I like most of what the Honda can do. I like the phone integration and that I can toss it down on a pad and it will charge while I drive. I like that the wipers, headlights and high beams all activate when needed without input from me. I like that it remembers my seat and mirror settings and restores them when I enter the car. I like keyless entry and remote start. I like the backup camera. I like adaptive cruise control, collision mitigation braking, and lane keeping assist. I like torque vectoring all wheel drive. Heated and ventilated seats are awesome.
Of course, if I were to try to keep this car for 14 years I might regret it, I don't know. Honda makes good cars but there's a boatload of stuff to break, lol. I think I might be better advised to trade it in at five years or so, but I do like the tricks it can do.
I'm not sure that the difficulty of performing a trick is related to its utility, and it doesn't really make sense to compare the usefulness of high beams and foglamps since they are indicated in completely different situations. Foglamps are mounted low and are usually yellow and meant to illuminate the road surface in heavy fog. High beams are counterproductive in fog.
I drive an old car partly for this reason. I've never had a car younger than 10 years old. My previous reasoning when I was younger was that I don't want to deal with the inspection requirements that come with OBD2 cars. Now I'm not so much worried about that with all the extra nonsense they're putting into them. My favorite car ever was older than I was, I loved it, an old ramcharger with power windows.
My current one has some "modern" archaic crap in it that was all the rage when it was built, it's GM so it has the OnStar buttons in the rear view mirror that are useless, along with complementary antenna on the roof. It's stock stereo controlled the alarm, it has all sorts of compartments that are nowadays mostly useless and IMO just an excuse to cram something in every spot. A lot of the "features" are just dead fads. The seats have lumbar shit in them and warm up. The warmer is useful in the winter I guess. Still with the bloat it is significantly better than the Frankensteins that commercials have convinced people they need nowadays.
My next car is either going to be a ~1995 4runner or a ~1994 diesel f-350 that I build myself. I do miss the corner windows.
I'm definitely in the same boat wrt complexity. My pickup has just enough extra electronics to be a quality of life improvement over a model that lacked those features. It doesn't try to drive for me and second guess my actions in unexpected ways. It doesn't give me unfathomable warning beeps that only serve as a distraction trying to figure out what's causing the beep.
At the same time it's important to recognize there's a difference between software bloat and growth. Security fixes often cause code to grow (checks, verifications, etc). That growth isn't bloat. The previous version of the software was exploitable because it lacked the checks that were added. Adding drivers or better handling edge cases in drivers grows code but isn't bloat.
Even "bloat" that's only added on-disk size (a secret Tetris game in some code) that doesn't affect normal code flow isn't the same as bloat as adding some advertising telemetry in the middle of a critical code path.
Not all code growth is bloat and not all bloat is equal.
I don't mind complexity but what I can't stand is having to access that complexity through touchscreens. Without the tactile feedback of a knob or button I have to take my attention off the road to see where I need to press on the screen. That's a major step backwards in auto safety, IMO.
And don't get me started on the lazy manufacturer design trend of bolting a tablet to the dashboard and calling it a day.
I am old enough to remember "Gorilla arm syndrome" talks.
People just love these big tablets even tough they are an ergonomical non-sense and they have always been for decades. Resistance is futile. Voting with your wallet is useless unless you enjoy the hermit lifestyle.
Mazda would make a great EV, but that is sadly not on their roadmap for the foreseeable future. I have been told this is probably an influence from Toyota's large share ownership, not wanting competition in that space.
That's not smart by Toyota (unless they own a very small stake), is it? Mazda competing with Toyota is indeed bad for Toyota but it's also bad for all the other auto manufacturers too. The benefit of that competition goes exclusively to Toyota (among other shareholders) but the cost is shared among all automakers roughly in proportion to their EV market share.
Mazda has been focused on continuing to improve internal combustion engines over the past couple of decades.
Specifically, the "Skyactiv" technology they developed was built to help improve efficiency and emissions of modern ICE engines. It's likely Toyota wanted to have some group/company (like Mazda) continue working on improving ICE engines while also investing in EVs.
I think ICE vehicles will continue to have a market, especially in very remote areas of the world where power grids are non-existent or not very reliable.
MX-30, Mazda's first EV but also ICE model, features touch panel A/C control for unknown reason. I tried it and found that it's difficult to adjust while driving. The decision really weird since Mazda avoid touchscreen for navigation and also avoid for A/C on other cars.
Mazda hasn't sold a production car with a rotary engine in a while, I believe since the RX-8 went out of production. Due to the nature of rotary engines they inherently have worse emissions (more like a 2-stroke than a 4-stroke) and while people have been hoping for another production rotary from Mazda it's unclear if they will build another one or not.
I'm kind of hoping the LiquidPiston design works out, since in theory it should fix a lot of the fuel economy and emissions issues (combustion chamber is closer to the ideal round shape), and maybe even be lower maintenance (apex seals move to the non-moving part of the engine, where they can be lubricated more easily). It's not owned by Mazda, though, so even if it works it's anyone's guess if Mazda would even be able or want to license it. Or if gas engines in general will be effectively obsolete by the time the technology matures.
I love the idea of rotary engines, but environmentally they're a disaster (at least in their current form), and even if the problems are fixed they'll still not be significantly cleaner than any other gas engine.
(I'm currently working on trying to covert an RX-8 to electric.)
I agree with you about wanting a car without pointless stuff bolted on. When it comes to the things required to actually move the car, there's an interesting inverse relationship between visible levers and internal complexity though. Each lever removed is moving complexity from your brain to some physical system. Like an automatic transmission removes the gear shift and adds the more complex automatic transmission. In the limit, one can imagine the "simplest" interface of a virtually empty self-driving pod, which of course is actually an extremely complex system.
Being as good as the desktop experience would be amazing. Realistically it will be much, much worse than the desktop experience because the auto companies are far, far worse at software than even 1990s Microsoft. You are talking about an industry where OTA updates are considered a major and challenging technology and multiple instances of bricking have occurred.
I'm with you. I will not buy a vehicle of any kind that is sending telemetry or tracking without my express written consent. This includes ICE cars made after 2018 that may have ODB3 and send GPS, emission, speed and other data to 3rd parties. I accept that I will be paying a premium to keep older vehicles running. I am about to donate my old truck to a charity and will get a less old truck after I leave California.
In fact all the links to forums discussing this appear to have been wiped from Google and Bing. Perhaps this topic is off limits for now. I can still find the older articles talking about privacy concerns, but the forums where people were explaining how to disable it have vanished.
' For all of us who feel only the deepest love and affection for the way computers have enhanced our lives, read on. At a recent computer expo (COMDEX), Bill Gates reportedly compared the computer industry with the auto industry and stated, "If GM had kept up with technology like the computer industry has, we would all be driving $25.00 cars that got 1,000 miles to the gallon."
In response to Bill's comments, General Motors issued a press release stating: If GM had developed technology like Microsoft, we would all be driving cars with the following characteristics:
1. For no reason whatsoever, your car would crash twice a day.
2. Every time they repainted the lines in the road, you would have to buy a new car.
3. Occasionally your car would die on the freeway for no reason. You would have to pull to the side of the road, close all of the windows, shut off the car, restart it, and reopen the windows before you could continue.
For some reason you would simply accept this.
4. Occasionally, executing a maneuver such as a left turn would cause your car to shut down and refuse to restart, in which case you would have to reinstall the engine.
5. Macintosh would make a car that was powered by the sun, was reliable, five times as fast and twice as easy to drive - but would run on only five percent of the roads.
6. The oil, water temperature, and alternator warning lights would all be replaced by a single "This Car Has Performed An Illegal Operation" warning light.
7. The airbag system would ask "Are you sure?" before deploying.
8. Occasionally, for no reason whatsoever, your car would lock you out and refuse to let you in until you simultaneously lifted the door handle, turned the key and grabbed hold of the radio antenna.
9. Every time a new car was introduced car buyers would have to learn how to drive all over again because none of the controls would operate in the same manner as the old car.
10. You'd have to press the "Start" button to turn the engine off."
Bollinger is extremely cool and I watched them with great interest for a while, but their initial estimates of ~$60k turned into $125,000 along with the short highway range puts it solidly into the "wealthy person's weekend toy" category.
As a software developer, I'll take a car without a computer, or with as little computing as possible any day! I do understand that fuel injection and abs are great stuff, but those could be ultra low cost asics, and, frankly, computing need not apply anywhere else.
In a parking spot, I never turn the steering wheel without the car moving forward or backwards. It's bad for the tires. But what drove the lesson home for me (pardon the pun) is when I drove a rental car with no power steering, sometime in 1990 in Europe.
If the car is not moving, and you try to turn the wheels with no power steering, you feel massive resistance from the road-rubber interface. When you let the car move this way or that, even just a little bit, that resistance dramatically eases.
To hell with power steering; what's it good for?
Power steering and brakes are dangerous because they are powered from the engine. In an emergency situation, exactly when you need the controls to respond in the expected way, your engine may die. Suddenly, you have no braking power, and the steering is all weird.
> Suddenly, you have no braking power, and the steering is all weird.
One day in in the early 1990's sometime, I was driving a car that started to aquaplane on a patch of wet road. Suddenly the car felt like a boat, and shimmied a bit and spun around. It caught some traction coming out of the wet patch, and at that point I could have stopped it faster, had the power brakes not failed respond in the normal way because the engine cut.
So I ended up striking the concrete divider with the rear corner of the car. I was not moving fast at that point; I didn't hit it very hard, but enough to cause damage and that could have been avoided had the brakes worked.
The engine might have cut because of lock up due to braking without traction. This was an automatic transmission, so that's another piece of idiocy to blame. In a manual, you wouldn't brake with the intent to stop without also hitting the clutch to disengage the engine; there is a good chance the engine and power brakes would not have cut out if that had happened in a manual.
(Electric cars should also fix all this.)
It could have been a pile up; this was in fairly busy, fast moving traffic with cars all around. Everyone behind managed to avoid me, luckily, and, equally luckily, the cars in front who were likely oblivious to the situation did not come to an unrelated halt, which would have seen me careen into them.
Well that's two different things. Intrinsic versus extrinsic complexity.
The user experience may get simpler again, but the technology inside is likely to get more complex. Electrification might mean simpler mechanisms and fewer moving parts, but the software will get ever more complicated.
Now in a way it's usually a good thing, if all the complexity of something is hidden and users can treat it as though it's simple. But we'd probably all agree that extra complexity in software that can kill us if it goes wrong is worrying. The only way we know to write safe software is to make it as simple as possible, and write it slowly and expensively. SIL-rated software has already reached the automotive sector. But it seems like the sheer demand to make cars more complex (especially for self-driving) will outrun our ability to make them safe.
Cars like Microsoft Word have a broad feature set because needs are broad. For those of us who want reliable, inexpensive, self service-able transportation there are fewer choices. My guess is because the majority are entranced by sexier things: smart features, driver assist, and safety features of dubious quality.
I used to live in NYC and bought a disposable car. It was a used rental car that was rear ended, but otherwise brand new with less than 1000 miles on it. I owned it for 5 years with zero issues and I just traded it in for almost the same price I paid for it.
I'm with you on this one. I think perhaps my favorite car interface to date is my '01 Camry. Let's look at the HVAC system:
Three knobs: temperature, speed, vent selection.
Three buttons: a/c, recirc, rear defroster.
That's it. The entire system. It's a 20 year old design at this point. I have never once thought that I, a mere mortal, could improve on the design in any meaningful way.
The buttons and knobs are big enough that I can work them w/ my huge gloves on, I do not need special gloves that work w/ a touchscreen, the buttons are not context sensitive, and I can hit all the controls in the dark without even looking. Another point, which I've now learned I had taken for granted, is that no software update will ever mess w/ my muscle memory. Those switches cannot be programmed to do anything else in software, as they're physically wired to the functions they control.
The beauty of the system above is that the physical controls are the state machine. I can literally feel the state the HVAC system is in. Nowadays that state machine is all in software, and I have to hunt through menus, look at a display, sometimes multiple displays, just to figure out what state I left the system in. That's just unacceptable.
I once drove a Volvo from the late 1970s with almost the same interface except one better: the air conditioning was a knob rather than a button, so I could select how "hard" the compressor functioned independently of the thermostat slider for the heater core and the slider for air routing. It meant I could dial in how much dehumidification I wanted while running heat in the winter and using recirculate to keep out some of the nasty truck exhaust and dust from sanded mountain roads.
I'm pretty sure the TJ-era wranglers (1997-2006), and the Cherokee of the same period hit a sweet spot. Dead simple, no frills, durable as hell, utilitarian but not uncomfortable. Everything after that is continual bloat - bigger overall size, more luxury options, more technology in general. But that's what people want - luxury, comfort, safety... I get it, but I love the experience of my TJ.
One of our cars is a 2004 Toyota Tacoma with everything analog except the radio, and I vastly prefer it. If it just had a usb plug so I could play music from my phone, it would be ideal. Oh, and oddly, the radio doesn't have a clock. Bizarre.
It I can 100% everything better than our 2016 Subaru Outback with it's annoying touch screen console.
Plus my nieces and nephews are amused by the actual manual window openers.
In my recent Subaru Impreza, CarPlay works just fine, however when I start the engine and put it in reverse, it will mostly ignore the volume control until you finish backing up and then a few second after you shift into drive/forward, it will then process the volume up/down messages it received in the meantime. It's like it dedicates all processing to booting up the screen and then in displaying the rear-view/backup camera. Also, it defaults to playing the radio (even if you had it OFF before, I think), so if you had the volume up for a quiet podcast or something it will blast the radio which is quite annoying, especially if you're in the middle of a phone call or something. I want a physical potentiometer which directly controls the amplifier to the speakers, like in any normal old radio. (This is also one of the exhibits in my mental list of reasons that people should write all the software they run on their devices. Or at least be able to.)
I don't think it's bogged down processing. It is explicitly muting the audio as part of the reversing protocol to avoid distractions, and possibly to let you hear other sounds such as from a parking assistance system.
My '12 Tundra had the backup screen in the rearview mirror. A huge innovation to me at the time. Everything else was knobs and buttons on the dash like yesteryear. That truck was awesome. A shame I had to give it up!
As an embedded software person who is also into cars, I am definitely going to be looking for something older for my next car. I have the luxury of not needing to drive it every day, but I like the idea of something I have at least a hope of repairing myself when it breaks.
Hey I'm with you. Not just in cars but for gadgets around the home. Coffee makers, microwaves, fridges, thermostats. Please give me as few moving parts as possible, the consequences of decisions made outside my control are a huge unknown with potentially large impacts.
Digital components may well be more reliable than corresponding analog ones. They can also be purely hw, with no sw or just a simple firmware. There is a middle ground between analog circuits and a general purpose programmable computer with windows 10 on it. :)
The biggest thing for me is user input latency. It doesnt matter if its a computer, a microwave, or a car. I want to feel like the machine is not a lazy piece of shit and actually wants to help me.
I know it sounds like a pedantic annoyance, but that little bit of step-wise discrete behavior I get out of my electronic throttle body right at the threshold of activation is one of the most infuriating things about owning an otherwise "sporty" car. It's not defective either. This is the cost of doing business with a totally-unnecessary software control loop.
I find that mechanical linkages usually have zero fucking latency, infinite resolution, and are much preferable to my monkey brain. Fly-by-wire is a huge mistake.
Are you sure that's not something else like inertia in various mechanical parts or engine tuning? Maybe it sacrifices responsiveness at low RPM for better emissions or better performance at high revs. There's certainly no need for the software side of things to be laggy or discrete.
Also have you tried an electric car? You might be impressed. EVs may have more software, but they don't have nearly as many physical constraints on responsiveness. They don't have to wait for an engine to suck in more air. They don't have to overcome the inertia of pistons and rods and flywheels and clutch discs and long driveshafts. It's just instant torque. Compared to my Model 3, every internal combustion vehicle feels like it has turbo lag.
You can't have high tech safety features without high tech, but you can limit the complexity of electronics, software, and in-car networking to a bare minimum, which is arguably not even being attempted, at least in some markets.
I wouldn't want anything resembling a golf cart, but in principle yes! Tesla has an OK idea of minimal dashboard, but then a less OK idea of tons of features on the screen. I don't know if automaticity or app based control or minimized function is the answer, but less is more to my eyes.
Manual transmission is simple and requires little to no software, but it's not simple for the end-user. Automatic transmission is simple for the end-user but involves a lot of software behind to make it work smoothly.
I think it's a damn shame that the opportunity for truly simple cars isn't occurring with the advent of mass market EVs. I suppose that ever-increasing safety concerns will push for cars that are more complex.
There's some amount of entitlement in that exclamation, though. The world, I feel, does not actually owe you to make any sense, or move in any predictable, safe pattern. The world will shift, and leave you behind.
Feeling resentful about that is just unproductive. (To take a slightly trolling tone, can you imagine how a lot of racists (BLM) must feel? Or religious fundamentalists, homophobes feel (gay marriage). The world does not owe them anything.)
> Damn. I just want a car with as FEW knobs/buttons/levers as necessary.
In 2005, there was a huge publicity campaign to make cameras mandatory in cars on the basis motivated by some parents driving over their toddlers playing in the driveway.
Of course, there are many rather more straightforward solutions to that problem. Given the fact that such parents must be in a rather select group, I doubted having cameras would prevent other sources of injury to their children even if they prevented parents from moving them down.
Now, auto manufacturers had to put a camera in every car. Yes, the requirement took effect in 2018, but the manufacturers knew it was coming.
So, you have a screen in the car. In addition, there is the temptation to monitor and collect information. You can sell that to insurers as "anonymized" data. Etc etc. You can enhance the tracking dimensions a lot if you can also get to track all the other stuff people do. Offer them the honey of pairing their phones with the entertainment system, and, boom.
> 221 people were killed by non-traffic (not on public roads) backover crashes in 2007, and 14,000 people were injured.
Given that the value of a statistical life is about $10 million, if the backup camera eliminates all such deaths and injuries, its benefit is about $5 billion. Conveniently, we do not know how many fatalities and injuries have been caused by trying to change stations or select tracks on a touch screen. So, on the cost side, we are stuck with the cost of all the equipment, software development, bug tracking, and, above it all, the annoyance of having to live with all this additional equipment one cannot avoid for the foreseeable future.
Nothing worse than automotive software. Buggy, slow, terrible user interfaces, outright dangerous and in many ways much worse than the systems they replace or augment.
The automotive industry has a long long way to come - assuming it will happen at all - before they can be said to be responsible software vendors.
Case in point: my - former - C class Mercedes that made two pretty good attempts to kill me by slamming on the brakes in a situation where that was totally unexpected and caused a perfectly safe situation to turn into a critical one. If not for playing ping pong for many years I highly doubt I would be writing this. After the first instance I had the whole car checked out to see if there was any fault in the system, the answer was that it was all working perfectly (that time the car had braked whilst on a very narrow bridge sending the car into a skid which I managed to correct before going over the side). Three weeks later it did it again, this time apparently because an advertising sign in a turn generated such a strong radar return that the car thought I was about to have a frontal collision. Again, out of nowhere an emergency stop.
I sold the car and got one where the most complex piece of software is the aftermarket radio, it has ABS and an ignition control computer but nothing in the way of 'advanced safety features'.
My vehicle actively trying to kill me is something I can do without.
So: as far as I'm concerned much less software on board of cars, open source it all if possible and roll it out much slower so we can get the bugs out.
I’ll pile on with my own anecdote. Last winter I flew home to Maine to visit family. Rented a ‘21 Nissan Altima. It drove well until there was typical New England snow/slush mix.
While driving on a flat straight stretch of road the car suddenly... yanked itself sideways. Thank god no oncoming traffic was present and I was able to course correct safely. I immediately drove home and paid careful attention to the wheel response. It kept feeling like it wanted to yank me off the road.
Once home i broke open the manual and found 4 different “driver assist” and “driver comfort” functions. After disabling them all the terrifying behavior ceased.
I’ve lived my whole life in Maine, Rochester NY and Colorado. I’ve never felt as unsafe in a car as I did with those software features enabled in about an inch of snow.
Bonus, it also has collision detection warnings on the side of the car. It was convinced every puddle I drove through that splashed slush beside the car was an object I was about to collide with.
> Bonus, it also has collision detection warnings on the side of the car. It was convinced every puddle I drove through that splashed slush beside the car was an object I was about to collide with.
That's the problem I have with people who tries to rethink our relationship with cars; it's obvious that they live a Californian lifestyle. "Just share your car" "Electric and solar is the way" "AI safety with cameras is a must"
The user interface of a car was standardized 100 years ago. If you need to read the manual of a car to decide which crap feature to disable seven menu layers deep in a configuration menu - assuming that the function to disable it even exists - then that vehicle should come with a mandatory examination before you can drive it and should not be allowed to be used for rental / short lease fleets.
Your typical manufacturer makes a very large fraction of their income from exactly those two applications.
If there is a mindset that has to change then it is that sticking something somewhere in a manual is sufficient disclaimer to be able to claim 'user error' in case of trouble.
Strongly disagree. If you mimic the established user interface of the car people have been using for generations, then it should behave like people expect it too. If special qualifications are required to drive it, then you should sell it only to those who can prove them.
You cannot build something that looks like an elevator, has buttons with numeric labels, and when people push one button the doors close and they find out they just selected the power setting of a giant microwave oven. I mean, you can build it, if you don't care about people getting microwaved.
Maybe the problem is, at best, alpha level software is being field tested by million’s of people with life and death consequences. Cars computer systems should always fail safely. If that can’t be designed right now then test it with professional drivers.
One of the main problems is that we have many different software teams trying to solve the same safety feature driven functions in many different ways with many different solutions, outcomes, and decisions along the way. This leads to a very fragmented base of software with varying levels of safety, none of which is tested to the same standards. The result is that consumers don't know of the car they are buying is truly safe at all
This is something agencies like NHTSA and/or private companies like IIHS are supposed to solve, but most active safety technology testing is just ‘does it work’ in very easy scenarios - the test suite needs to include many environments and tough conditions if it wants to evaluate these systems more broadly (The only issue being when car companies optimize for those specific tests, or ‘when the measure becomes a target’).
Anyone want to fund me? I’ll test the every living heck out every smart car in existence in every condition from perfectly flat straight drag steps to dirt roads in the mountains with a foot of snow and white out conditions, to unmarked pot holes riddled roads in torrential down pours. I’ll also test the usual stuff like does the car detect the kids playing in traffic, or the dog darting after the squirrel. I’ll video the test from numerous angles, cut in awesome music and sounds effects and upload everything to the internet. I’ll also write long articles about the intricate nuances of each vehicles responses to the tests and create scoreboards for categories like least likely to run over children, or most likely careen off a bridge if there’s a sign at the perfect angle.
It couldn't be disabled either without breaking out the wire snips and selectively disabling stuff and hoping that that would not have further averse affects. That car had a radar unit behind the front license plate and a camera mounted in the windshield. No idea which of the two was responsible for the false triggers.
I would assume incompetence before malice and living in some kind of idealized bubble compared to field testing and dog fooding their own product in a very large variety of circumstances.
One car executive that I've known said this: "The very best test drivers are our end users, in the first week after release we learn more about a car than we do in the months of field testing prior.". I am not going to name the individual but it makes good sense and I have seen the same with all software products released to the general public.
I work in the field and never had this impression. There are a ton of standards, safety related procedures and so on. But many of those are at least questionable. At the same time management is asking for more and more „features“. Many are just gimmicks or diluted by too many cooks. No one ever asks for GOOD software. Folks at the top have ABSOLUTLY NO idea what that could mean other than the absence of bugs.
I bought a car with an aftermarket radio that has a DVD player in it, which I've never used. When you turn it on, it displays a message warning you not to watch a DVD while driving. Sure, whatever. While the warning message is up, you cannot adjust the volume. I have to make sure to turn down the stereo before parking otherwise there's a solid five seconds on startup when it can't be turned down. Insane.
I have this old aftermarket $30 brand new at the time crappy little stereo. No optical drive, the thing doesn't even have an AM receiver. But it does take SD cards, USB drives and AUX. Perfect. That's exactly all an aftermarket system should do, maybe Bluetooth, but I don't use Bluetooth so I don't care. I can fit ~60 albums on one 8gb SD card. Goodbye buffering in the boonies, goodbye book of CDs.
There's no way to navigate directories besides just skipping to the next one. Fine. I've seen some that let you open a menu and navigate using the volume knob, but no SD cards. I tried very hard to find one that has an SD slot and more advanced directory navigation. The two are apparently mutually exclusive. So instead of getting to decide what to listen to, my unit mostly functions as a commercial free music only radio station with a wide, diverse selection of exactly things I like. I can deal with that.
It doesn't load the directories or files in any particular order, like it builds an index by randomly loading directories as they're read and then the files inside. At least it doesn't mix up files between directories. I can deal with albums being out of order I guess. But why it doesn't just load them in alphabetical order is beyond me.
Nothing but MP3 support. I even wrote a bash script to turn FLAC files into MP3 files just for the car, and it can even split FLAC by CUE file. Works just fine for me.
The thing has a predetermined EQ animation. It does not actually correspond to the output, its just an animation. Why? Who knows. But it does.
All in all I've been using the same unit for 5 years across 2 different vehicles. I like it I guess. The window smashers seem to know it isn't worth anything, which is great. If I could navigate better, it loaded directories and files alphabetically, and had an AM receiver, it would be the last aftermarket system I buy for the rest of my life.
My suggestion for getting SD card support is using a USB adapter. I'm using a Sandisk one . It's quite small and hardly sticking out at all, and I just have it plugged in to the radio permanently.
Otherwise, my Pioneer DEH-X6650BT let's you switch between directories with a button press. Basically you have previous and next track, as well as previous and next directory. The toplevel is simply another directory. It also remembers where it left off after you start it, and it is super snappy with 19GB of MP3s that I have in various directories. Also, the manual is very explicit on what file system is supported, how many files and directories and in which order files are being played. See this page and the next one . The only downside is that I can't seem to turn it off. As in it the radio is always illuminated, even when you are not playing anything.
Just leave the control to the person behind the wheel, but somehow that's too stretched of an idea.
I hope the F150 E truck will not have those 'automatic' features. but then again, I kind of wanted lane keeping and follow so maybe that's just the necessary evil. Maybe 'modes' where it gives all controls to me when I wanted it?
Counterpoint: My Subaru has averted at least 2 collisions till date by applying an emergency brake (admittedly not in any life threatening situation, though it did save me a considerable amount in potential repairs). Though yes I agree, the option to turn off features is certainly helpful.
You may want to go through some additional driver education, if your emergency braking system activates several time in a short period of time that is a possible indication that you may not be on the ball. But happy to see that you ended up avoiding two accidents :)
lane keeping is making me dizzy, at least on the rav4 where I had it. It's a constant battle between me and the computer about where to keep the car in the lane, making it such that the car constantly sways: 1. corrected by me, 2. then corrected by the computer, goto 1.
If I don't correct at all, the system is getting mad at me for not keeping the hand on the steering wheel.
Auto follow is again somewhat risky, you can't really rely on it, sometimes it brakes too late.
Serious question - what alternatives are left for those of us who want a dumb car? I've spent the last three months finding out I can't get solar panels installed without a high-fidelity power-monitor tap connected to the provider's cloud, logging every appliance I use and what it's doing. Same deal with cars - they are on the internet and they generate evidence used to convict, and geofencing is coming. Other than stockpiling cars from 2010 - will we have alternatives?
The vast, vast majority of all new cars sold today either don't have any kind of uplink to the cloud, or can have that capability easily removed (e.g. OnStar from GM). It's really only some of the EVs (the non-compliance ones like Tesla, really) that are software-heavy and blazing a new anti-privacy trail.
There is an extremely limited amount of diagnostic data recorded by most cars (Tesla aside) - absolute limits, performance counters, and freeze-frame diagnostic data. There simply isn't much storage, and again, Tesla aside, most manufacturers don't want to have to buy high-write capable flash of the sort that could cope with constant logging.
The most invasive is probably airbag blackbox data, which is stored upon deployment and not routinely uploaded besides as part of an investigation.
As far as I know based on extensive reverse engineering of many modern European vehicles, no location data is routinely stored or uploaded to a dealership tool by any of them.
As a car enthusiast this sounds laughably new to me. My main vehicle rolled off the assembly line in 1991. Unfortunately it's a "hobby" and a labor of love because if you're not wealthy enough to pay for a shop to handle maintenance of your classic cars then you're spending your weekends working on them.
I get the allure of older cars but personally I wouldn’t ever daily drive a car that didn’t have vaguely modern crashworthiness. To me 2010 sounds like an ideal middle ground between maximising mechanical safety and minimising the encroachment of bad software. Ideally you’d want a reasonably new safety-rated car where the entire radio can be swapped out for an aftermarket double-DIN unit.
I don't think there are any. If you got to the point where you somehow managed to source the labor, materials, and space to manufacture "dumb technology", you'd be sued into oblivion by competitors who wouldn't want you to eat into their profit margins.
I think the market is huge. Imagine a company that suddenly started selling all-metal consumer appliances with minimal functionality, controlled by old-fashioned switches, buttons, and knobs, designed to be repaired. They'd be insanely popular. Of course, that lack of subscription-model pricing and the high labor costs of worthwhile designers and manufacturers would also destroy the company, but my god, a boy can dream.
the problem is that most of these are horrendously oversized for private use. So I would like to have a commercial agitator to create yeast dough; these aren't even that expensive for their small series and local production - compared to e.g. kitchen-aid stuff -, but I both don't have the space/floor to put down a 4sqft/1000 pound appliance and I certainly don't need a 5 gallon bowl for my dough... I'm sure you could create and sell a home-sized variant of these, but yeah, distribution won't be easy.
I have to admit though that I sometimes dream of setting up a company, which just goes through all these relatively low-tech, electromechanical things and cyclically produces these. Problem is: it's capital intensive, low-margin, low-growth and requires you to keep people around. At this point you've lost everyone nowadays even if it is sustainable (nothing goes to waste when making these repairable) and keeps skills alive (which both the US and EU are paying extreme prices for through the military acquisition procedures).
In this vein I also suspect that most people could easily afford to get their whole furniture sourced in a "raw wood" edition (to be painted/oiled) from local woodworking shops, as the price difference is virtually nonexistent (built a bed with the help of friendly non-CNC-shops: was 12h of work, which clocks in at maybe 900USD. Material was 300.). If I go to a non-ikea, but industrial/imported furniture-store, similar quality would have cost me ~4000USD. But of course, a healthy ecosystem of woodworkers and designers would not concentrate wealth.
Note that commercial-grade appliances may have unexpected side-effects. I have a speed queen commercial coin-op washer and it's highly reliable and easy to work on, but you can't do anything but a full cycle on it.
Won't t 3D-printing metal appliances be more and more widely feasible in coming years? And, at least here in the EU, right-to-repair  labeling gives consumers the ability to more efficiently vote with their wallets. Its also happening in the US 
Liquid steel requires enormously high temperature and magnesium/aluminium require an argon atmosphere. Titanium requires both. So unless you're going to make everything out of zinc (assuming we don't run out of zinc!), this may not be the cheap fix you're hoping for -- and zinc isn't very strong.
There are also many types of steel that have widely varying properties.
There are some (expensive) printers that print powdered metals (including steel) that is later sintered in an oven. They're certainly not large enough to make a car, and even if they were the properties of the material are likely not ideal, and the process would be prohibitively expensive.
I've been using mine for many years. The only big problems I've had is that sometimes the scale doesn't work perfectly and I have to reset the weighting and that after so long the blades have been getting a little blunt; but they can be sharpened.
I don't think I'm willing to go that far. What constitutes 'the HN crowd'? The folks who come here and read the stories, but do not read the comments? Or just the ones who read the comments? Or only the subset of those who read the comments who actually bother to vote. Or participate with their own commentary?
It's pretty easy for a small subset of individuals to appear as if they accurately represent the wider group.
I found the latest model of the Suzuki Jimny the closest to what you described, that is when it comes to new cars. Unfortunately it has been withdrawn from the European market (where I live) because it doesn’t meet environmental standards, hopes are that it will be re-introduced labeled as a “utility” vehicle.
Afaik it is selling like hot-cakes on the markets where it is still present (like Australia), maybe if enough future potential customers ask for it Suzuki will decide to also sell it on the US market.
I bought a new Nissan Kicks last year (the lowest-end crossover SUV) and it seems pretty dumb. It has no internet connectivity as far as I can tell. It has a touch-screen, but it's pretty much just for the backup camera and audio controls, everything else has physical controls. It does have a computer that is involved in operating the CVT, but besides that it never gets in my way. And it does have sensors for the safety features (blind spot warning, etc) but that stays out of my way too. (The one annoying exception was when I had a bike rack on the back. I couldn't drive in reverse, the collision detection kept sensing the rack as an obstacle and slamming the brakes.)
Mostly-dumb cars still exist on the lower end of the manufacturers' lines. But yeah, who knows how long that situation will last, or if you'll be able to get anything technologically dumb with premium power and handling.
> I bought a new Nissan Kicks last year (the lowest-end crossover SUV) and it seems pretty dumb.
Not sure about 2020, but 2021 has Automatic breaking, pedestrian detection, and collision detection standard. There are likely a few hundred thousand to millions lines of code in these various systems. Even "dumb" cars today have tons of software, it's just hidden since it doesn't require user input. Literally every aspect of your driving is fully computerized - braking, acceleration, engine spark plug ignition, transmission, steering, etc. The infotainment is often relatively simple compared to all of the other software running internally.
I just bought a 2018 Subaru Crosstrek that's surprisingly dumb. All mechanical controls and a basic infotainment system that isn't required for the car to operate. AFAIK it doesn't have any cloud capabilities and isn't connected to the internet. I imagine the latest models are the same since they look similar inside.
EVs are quite simply and there has been a small, but growing base of users using AC induction motors on sailboats. I imagine you're going to need a pre-OBDII car and some mechanical skills and convert it to electric drive. After that you'll have a car that has half the range of a modern EV do to weight optimization.
I've got a 1985 Jeep CJ-7 sitting in a shed in another state right now. The 4-cylinder engine sucked when it was new, and 35 years of entropy have not been kind to it, especially the labyrinthine emissions control systems. But I'm holding on to it because I think it would be an absolute hoot converted to electric... plus, it's got such a small gas tank and bad gas mileage that a 150 mile electric range would be no worse.
You're late to the Land Cruiser party. Prices for FJ40 went through the roof a few years ago, then FJ60 started to climb, then even 80 and 100 series jumped in value. During the pandemic everyone wanted an overland rig and prices have really shot up in the past year.
I'm amazed that in this entire article Autosar wasn't mentioned once. The giant 2 ton elephant in the room here is automotives reliance on god-awful "kitchen sink" style standards. Try reading through the various Autosar docs and ask yourself if you expect robust bug-free code to be written to comply with it.
There needs to be a complete cleaning-of-house in automotive software.
I2C, Flexray, Ethernet, CAN/CANFD & OBD, LIN, what are we even doing?
ARXML, FIBEX, DBC, fuckin kill me.
"Unmanaged complexity" a.k.a "we've never thrown away a single technology or standard even once"
I worked in the automotive industry and I can totally confirm. The worse thing is, there are now AUTOSAR experts and AUTOSAR tools and AUTOSAR Tool Experts and within that tool there’s a ARXML generator that’s generated with another tool …
There’s no way this can ever become safe, robust, software. The worse part is, there’s so many careers that depend on this obscure skillset that I am unsure a change can come from existing companies.
Dont even get me started on the tooling. Vectors, Elektrobits, Dassaults, Conti's, tools are probably one of the biggest drains on collective computation power outside of crypto and ML.
Not to mention working in this space is fucking soul consuming. I was considered an "AUTOSAR expert" for a time, and that essentially meant having enough programming and systems knowledge to work on the entire stack. But never writing a single line of code, only clicking buttons in these god damn tools and watching them crash constantly, loosing hours upon hours of work
The problem with this is that auto companies are not software companies. They may have good engineers there, but they are hamstrung with a culture that considers software as an add on cost center at best.
Perfect example: I have no way to report software bugs to Honda. I've found a few and collected detailed reproduction data. The best I can do is give it to a sales rep in the service department and hope they send it "up to corporate".
Compare that to Telsa, which has bug reporting built right into the software in the car, as well as bug bounty program.
And then there are updates. Honda found a bug where the speedometer would just crash and not show your speed anymore. This is was pretty bad, but I had no idea about it until I went into the dealership. There was apparently a recall but I would have had to find that myself, I didn't get a notice. Honda has no built in facility to notify people of software updates and recalls. And then once I found out, the only way to fix it is for a dealership to apply the update. There is no over the air update and no way for me to apply it myself.
Car companies need to learn how to be software first, or things will get very dangerous.
On the other hand, having seen the software industry I don't really want its values applied to cars.
Games used to be shipped on chips to customers with the assumption that they could never be updated, but now the expectation that that's possible results in multi GB release day patches.
Games used to be shipped to consoles without network access but if Blizzard goes out of business tomorrow, Starcraft II will just fail to boot because the software you get isn't enough to run the game.
Games used to make most of their money by selling you a fun game that you wanted to play, but now they make most of their money from DLC's or selling in-game currency.
I don't want my car to move fast and break things. I don't want my car to fail to drive me to the hospital because it has updates to install. I don't want my car to drop features it had when I paid for it because some PM's bonus depends on me using their monetised features instead. I don't want my car to be measured by "works on my machine"-level QA. I don't want my car reporting telemetry. I don't want my car's UI to become unuseable over time because the developers shipping constant updates to it are working on the newer hardware instead of the hardware I have. I don't want parts of my car to stop working because an app developer doesn't support my car anymore. (I've lost access to many an iOS game because of this without changing any hardware myself.) I don't want my car to stop working because it doesn't support TLS 4.0 or 6g and the licence server requires it. I don't want my car to stop working because it can't find the licence server because Volvo went out of business, or even just decided they didn't want to support me anymore. I don't want my car to fail to unlock because of a network blip. I don't want my car to suggest I go to McDonald's instead because of an advertising deal.
That's what embracing the software industry's norms will get you. Awful QA, shortsighted dependencies, terrible incentives, and the ability to monetise you.
OK but with the complexity of games today there is no way such a game could be released bug-free. Could games be released better than they are though? Probably. We needed game software updates in order to allow the 10x complexity increase, but developers took advantage of that to also give them a couple more months of development time.
Not every company embraces those sorts of worst practices. Especially hardware companies. Tesla and Apple are good examples of this. They have strong QA policies and make their money from hardware so they don't have to do all those "you must be connected" tricks. Sure, both still have bugs, but not like web apps and games.
Most of what you listed applies to web apps and games, not bespoke software for hardware.
Maybe not a "dumb" car, but yes, I would like to see more good analog options if the digital alternative is getting screens in cars that look utterly embarrassing compared to yesteryear's netbook-sized screen fad.
If you're going to put tech in my car, you better go all the way. I'm talking a huge screen, fast multicore processor or redundant systems, touchscreen to UI update response times under 5ms.
None of this nonsense where you're getting some baby embedded system and the screen updates over 30-50! ms. Shame on these manufacturers. In 50 milliseconds at 65 miles per hour, I think you've moved like over 4 feet. That's ridiculous.
Say you've got an interaction that takes 150ms. At highway speeds you've moved the entire length of a car.
This stuff is simply unacceptable. I mean to the point where I want regulations on how slow your crap software can be. If I'm moving 4,000 lbs down the road, I don't want to be distracted. I want the exact same responsiveness as an analog physical switch or knob.
I have to admit I was unnerved the first time I got in a car with an electronic hand brake - certainly something I would not want if I were buying a car. Especially as there was a "Microsoft" logo next to it (presumably for the terrible in car entertainment system which was touch screen only)
The encroaching of software into cars does remind me of the old joke though.
At a recent computer expo (COMDEX), Bill Gates reportedly compared the computer industry with the auto industry and stated "if GM had kept up with the technology like the computer industry has, we would all be driving $25.00 cars that got 1,000 miles to the gallon."
In response to Bill's comments, General Motors issued the following press release -
If GM had developed technology like Microsoft, we would all be driving cars with the following characteristics -
1. For no reason whatsoever, your car would crash twice a day.
2. Every time they repainted the lines in the road, you would have to buy a new car.
3. Occasionally your car would die on the freeway for no reason. You would have to pull over to the side of the road, close all of the windows, shut off the car, restart it, and reopen the windows before you could continue. For some reason you would simply accept this.
4. Occasionally, executing a maneuver such as a left turn would cause your car to shut down and refuse to restart, in which case you would have to reinstall the engine.
5. Only one person at a time could use the car unless you bought "car NT", but then you would have to buy more seats.
6. Macintosh would make a car that was powered by the sun, was reliable, five times as fast and twice as easy to drive - but would only run on five percent of the roads.
7. The oil, water temperature, and alternator warning lights would all be replaced by a single "General Protection Fault" warning light.
8. Occasionally, for no reason whatsoever, your car would lock you out and refuse to let you in until you simultaneously lifted the door handle, turned the key and grabbed hold of the radio antenna.
9. Every time a new car was introduced car buyers would have to learn how to drive all over again because none of the controls would operate in the same manner as the old car.
10. You'd have to press the "Start" button to turn the engine off.
I 100% agree with this. I want climate control knobs and buttons, a basic AM/FM radio, and a CarPlay/android auto screen. Everything else is noise, and as Toyota would put it, MUDA (Waste / non-value added).
In my view, there is no reason for an auto manufacturer to invest heavily in their infotainment systems anymore - it just isn't a competitive advantage for most cars. Almost all users who buy the upgraded trims will have a smart phone.
I work in the automotive. I think that this is the trend. At the begin, smartphone projection was just calls, music and navigation. Now such systems are interested in signals coming from the car.. guess why!
150 million lines of code in a Ford F-150? How is that even possible? A Volvo with 100 million lines including 3 million functions. This sounds like generated code to me. I can’t believe this is handwritten or even necessary.
A bit of a naive question I admit, but how do you even test hundreds millions of lines of code to ensure they all fit perfectly when different ECUs have different suppliers, and by customizing the car you can have different types of chips? Just curious how integration of all the components is done.
Also, there seem to be a lot of recalls due to software issues, so I wonder if there's any open source or anything close to it that has tools such as CI or VCS for newer electric car companies that use ECUs from different OEMs?
I've been exploring building my own car from scrap. This wsj sort of motivated me (1). The ideal would be no electronics at all. An issue I have with new cars are monitors, I hate them they are distracting. My eyes are pretty sensitive to computer screens etc. Maybe, I'd settle for just a radio. "Maybe" because you then get looking at a cd player then all the sudden you want the further desire to control what you listen to and before you know it you are talking about mp3/digital and more computerization/softwaring of the car.
The first question is... what do I want? The second is the more complicated issue of getting it done. However it has always been a dream of mine since being a kid and watching the Home Improvement sitcom in the 90s.
It would be fun to go one step further, and turn a Tesla into a totally analog car. A Tesla might be a good starting point as they’re structurally some of the safest cars around (and this is one area where older, naturally analog cars compare badly) and turning the most technologically-laden car into the least... would be apposite.
Once again, a plea for moderation from those calling for a return to analog gauges. Remember that outside of Tesla, automakers are generally pretty conservative and most computing in cars outside the head unit is decentralized MCUs that don't connect to the internet. And within the head unit, CarPlay/Android Auto has moved most of the work to phones.
Peter Hubers tome The Bottomless Well considers software as the apex of the energy pyramid. Each level of the pyramid- animal, wood, coal, gas, electricity, nuclear, software- (I may have recollected the order not entirely correct) is more usable and powerful than the one below it.
You can see this pyramid in the evolution of the automobile: mostly petro-mechanical, then a growing fraction electrical, then an increasing fraction software.
I was not fully convinced by the book is that computing is a type of refined energy, but can agree with some of arguments for it. Other computer utilization like mass data centers and crypto currency support computing as the new wave of industrialization.
As an aside: Hubers thesis is the world will never run out of energy because we are constantly improving it, for example with or as software. Furthermore the amount of work per capita has grown with the quality of energy, and shall continue to increase in future.
In an electric car the motor control software can be quite tiny compared to traditional engine control. A lot of them are also direct drive, so no transmission controller.
Now battery charging is a bitch. The standard communication between a Level 2 charger and a vehicle is IMHO designed by committee. It uses power-line communication even though it's not over the high voltage/current wires in the cable. That means special chips, firmware, and TCP/IP. Sounds like a startup solution rather than just plain automotive CAN connection.
Anyway, most of the software isn't worse than an ICE car. Also, most of it will still be running on micro controllers, not fancy Linux systems. Detroit still knows how to do embedded but they're starting to get corrupted with ideas from all this autonomous stuff.
I remember back in the early days of Linux how we'd say that using a proprietary OS is like owning a car with the hood welded shut. It seemed so obviously ridiculous. Yet, that's basically where we're headed now.
In some ways that's a good thing: EVs require much less physical maintenance. (At least, their drivetrains need less maintenance.
Whether the rest of the car does depends on the manufacturer.)
But on the other hand, depending on how heavily locked-down the car is, it'll be hard to do third-party modifications and older vehicles are going to be at high risk of having security vulnerabilities as soon as software maintenance for old vehicles stops being a priority for the manufacturer.
What's the effective lifespan of a new car now? Software will likely shorten it in many ways. Especially with subscription models being so tempting for manufacturers. As the decades progress, will we care? But the street will find uses for these future old cars. Those powerpacks will be reused. Parts will be swapped.
Limiting factor: will manufacturers lock down components so they have a shorter and finite lifespan and/or compatibility? Very likely. The Right to Repair movement is still gaining traction. How this plays out will be quite a story.
I hope its not a repeat of most IoT where stuff frequently turns into junk within five years and usually heads off to some sad landfill.
With WFH one could expect the total miles driven per year to decrease drastically due to less commuting. This would make the lifetimes of cars longer and lead to lower demand -- a second way in which software is "eating" the car.
What if cars were like TVs? Some would be smart with integrated software, and some would be dumb and require a "stick" to make it smart. I'd certainly be tempted to buy the dumb version and have the flexibility to try different software experiences.
What worries me the most is the usage of flash storage backing huge swaths of the functionality in the car. I suspect 10 years from now we will have cars where everything between the dashboard and glove-box does not work because the flash has worn out.
Luckily the only dynamic part of most of these ECU's are their diagnostics and logging mechanisms, nowhere near "huge swaths of functionality". Every other bit of flash is written once at the factory and is never touched again. Now that we are seeing widespread adoption of OTA, it'll increase the amount of writes to program flash maybe 100 times over the life of the vehicle, but still within reasonable bounds.
The issue is if OEM's ignore this limitation and tie mission critical portions of their systems to the memory partitions their diagnostic and logging mechanisms use (e.g Tesla).
In reality if we are deploying OTA capabilities to these vehicles, there is absolutely no reason to be hammering your flash with logs, just upload them to the borg cube and be done with it.
Many cars use eMMC to back the infotainment system.
Recent non-tesla rentals I've driven have definitely had functionality that could only be accessed through the touch-screen. e.g. a Chrysler my mom rented a couple weeks ago had just temperature and fan control on knobs; everything else for climate was touch-screen only. What happens when the eMMC gives up the ghost in this car? I don't know, but I suspect you won't be able to e.g. manually put the defogger on...
Good point, you are totally right. I don't drive a car which uses the infotainment system for mission-critical functionality. I've actually run it without the infotainment system in the car at all. Tesla's would effectively be bricked without the head unit functioning.
Ford and the rest of the reputable OEMs know a thing or two about flash degradation though. It was actually a huge sticking point in one of my infotainment projects with them. So I don't expect all the OEMs to make the same bone-headed mistakes Tesla has
If you are allowed to share, what is the typical OEM specified lifetime for the infotainment flash part? I guessed 10 years in my original comment, but that was a gut instinct and completely uninformed.
I own Mitsubisi Pajero (Montero in the US) that is basically designed in the 80s-90s. I installed a dumb car radio with monochrome display, which plays sound via USB and BT. I want a car that is controlled by me directly: electrically, hydraulically and pneumatically rather than algorithmically. With only the exception for ABS and ignition control which are the only valid uses for the software in the car.
oh its funny.. of all the rust trend storm I cannot recall one mention of automotive industry usage (not that I imply it's not used.. I just don't remember hearing about that).. this is a place where I'd really love safer languages.
There's unfortunately no hardware or standards support for Rust, nor will there be any time soon. Silicon vendors don't provide Rust compatible toolchains, and the standards everyone is locked into is C/C++ only
I drive a 1985 VW Vanagon. While there have been many, many times that I have wished for some comprehensive diagnostics, and am considering an engine swap at some point, when I drive and there is literally zero distraction from electronics or touch screens, it’s wonderful. I did add a CarPlay-enabled deck for directions and road trip music, but otherwise everything is manual.
I have this weird feeling that we ought to displace the problem somewhat. I feel like the ubiquotous & pervasive computing people were onto something. And in some ways, we're already seeing a very narrow brand of this future arrive: Apple and Google both have systems to allow the phone to control & manipulate some of the car's infotainment systems.
Extending that idea further, & removing most of the native infotainment from the car, turning it into a bunch of dumb, wirelessly controlled displays & buttons, that an external system can use, would be interesting. Certainly there's still a large maintenance burden. And now we're talking about allowing external consumers of the car's services.
There is some precedent for this. Webinos was a very intersting ubiquotous computing platform, and one that BMW/Jaguar/Land Rover did a bunch of work on. It definitely still kept the car's infotainment system, but it also exposed many of the car's systems & services externally, over a normalized, secure, webinos control system, such that you could manipulate the car's systems, or in one demo, look at the radar system, from remote devices. I kind of picture the radicalized form of this as, your car has some hdmi ports in it, and you plug in a Roku or Chromecast or whatever to power the screens, or have your phone wirelessly send a video stream. The manufacturer would still need to have an out-of-box experience, but in 10 years or whatever, the manufacturer might not have to still support it like they do a built in one: they still have to maintain some API surface, but that, hopefully, can be a simpler, more controlled, known interface, with less maintanence burden, & less fancy application processors.
I don't really think what I suggest saves all that much trouble. It introduces more trouble too. But starting to decouple computers, starting to untangle the weave, but it does seem like a long term more sustainable course of action. Whatever modern computer we carry with us is what we trust, and leaving it to provide an up to date experience across all varieties of screens, inputs, peripherals we encounter has always been, to me, what the ubicomp revolution was about.
I've been shopping for new tractor and found out that some newer tractors have software-controlled regeneration process. Then I watched on Youtube how buggy software in those tractors randomly kicks in regen process and does not allow owner to use tractor. I ended up buying lower HP tractor that has no chip and no regen process.
I talked to quite a few car manufacturer in China:
As part of my research into automotive software and exploring opportunity for disruption.
They were all on the hype train of "software defined cars". Definitely not just a cost center. Of course they are still figuring out how to do software, and learning from people in the internet industry.
As someone in the industry of supplying parts to keep older cars running, I view the increase of automotive software and electronic complexity as ensuring a future crisis of maintainability.
Availability of parts and service information has always been an issue for aftermarket repair/modification of vehicles. However, as long as there are enough vehicles and committed owners around to create a small market for repair parts and services, independent companies have grown to provide what the original manufacturer will not. This even applies for very niche vehicles where some devoted old fellow runs essentially a hobby business keeps the flame alive.
Even relatively recent vehicles with considerable electronic sophistication can be supported this way. I have worked with specialist companies that will modify, repair and re-engineer some of the more complex control units and electronic subsystems used on 2000s and 2010s vehicles (i.e. suspension control ECUs, digital dashboards etc).
However, the trend described in this article has the potential to upend the status quo described above, simply due to the escalating complexity involved. It's sort of a tradition that auto enthusiasts and aftermarket industry initially distrust new tech in automobiles- fuel injection, ABS, traction control systems, emissions controls such as EGR etc all got that reception initially. Expertise with all those systems was eventually absorbed throughout the industry and resistance decreased as the benefits were better understood. However, as complexity increases there is a gradual increase in costs (engineering, training, manufacturing, install/service labor) to deal with all these sophisticated systems. Without other unforeseeable changes, there are almost certainly various inflection points where increases in complexity will result aftermarket support collapsing for particular models (or specific subsystems). This is something that already happens, but mostly for relatively rare models as until recently automotive complexity increases were constrained by the slower pace of ICE and chassis development.
As these costs rise, fewer and fewer models of cars will have a healthy enough aftermarket to support investment by independent companies to analyze, repair and replace these complex systems. For sure, it will result in more models of cars becoming unmaintainable and fewer cars staying in operation beyond their warranty expiration dates. However, also I think this will result in a market space opening up not for repairs and replacements, but for various kinds of bypasses and defeat devices ("deletes" in industry terms) that will either remove complex subsystems entirely or allow replacement with more generic components. This is already occurring in some sectors of the automotive industry, particularly around diesel truck emissions control systems, where EPA has pursued aggressive enforcement actions against companies selling delete kits. However I think where it will get really interesting is when we start getting widespread delete kits that aren't primarily mechanical in nature, but attempt to lock out or spoof entire software/electronic subsystems.
According to SEMA, in 2019 there was $890 Million in retail sales of parts for pre-1974 vehicles. Size of businesses encompasses everything from large operations like RockAuto or O'Reilly to small ultra-specialized one-man operations.
This makes me think of aeronautical industry. It is possible to get a J-3 Cub with just a stick that pulls on the control surfaces or it is possible to get a Cessna private jet with thousands (if not millions) of lines of code in it. They are just different.
My fear is that we'd get to a point in which dumb cars will be something that is no longer a mainstream option. Maybe then we'll need to have our own EAA but with cars.
Yes, something between a $500 bicycle and a $20,000 car. The barriers may be as much regulatory as anything else. In the US all cars must meet highway safety standards even if they will never be driven on the highway. I think it's even illegal to sell a car with a top speed of 35 mph.
There are some places, mostly retirement communities, where a lot of people get around by golf cart. They do the job and are a lot of fun.
Dacia has a new car in that bracket, but the range is a bit limited.
If you live in the US, Dacia is basically a Romanian company under Renaults ownership (which is a famous french brand) whith focus on cheap cars with proven tech. The use proven engines, parts from Renoult cars, etc.
A pretty good philosophy IMO but until now they had horrible designs, now they are improving, in the sense of adressing more general-public apetences, which means generic-looking SUVs.
There is Arcimoto, an operation based in Oregon. I think in some places their vehicles are grouped with motorcycles or something along those lines. I've seen them zipping around San Diego and other places.
Yea, but these cars aren't available or legal to drive in the USA. To import one, you need to pay fees and it must be registered. To get it registered, it must have safety standards on file with the federal government. Unless you want to pay the $$$ (millions?) to get that certification for your single imported car, you need to wait 25 years for it to be imported as a "classic" car so those safety regulations don't apply.
I would love to buy a Honda E, but I live in Texas and it will never happen. After 25 years, the tech will be so outdated that it's likely not worth it. Plus the batteries will likely need to be replaced and good luck sourcing parts for one stateside without paying $$$ in shipping again.
You can get an enclosed mobility scooter for $6k-7k. Many have some sort of stereo system and climate controls though I don't know if they all have your desired features. But they do exist and are for sale in the US and Canada.
The Boomerbuggy X! This fully enclosed mobility scooter is more spacious and has some cool new features! Travel to get your groceries, to your neighbour’s or take it just for a leisurely joy ride without fear of the weather. The Boomer X is fully insulated with heating giving you the warmth and comfort that you need on those cold winter days. The Boomer X also features built in speakers, windshield wipers, and more. Regain your mobility, independence, and sense of freedom with the Boomerbuggy X the next generation of covered mobility scooters!
This is true to some extent, but the biggest burden on automotive computer performance is actually safety compliance. Most modern automotive ECU's have an entire core dedicated to cross-checking the validity of the other cores execution. Protection against an extremely hostile EM environment ends up resulting in very low clock rates (150-300MHZ). And software safety mechanisms mean, at some times, 50% of core allocation could be consumed by safety related code.