Hacker News Reader

Comment by pwg | Hacker News Reader

> As a User Experience Professional, I was never able to grasp the true user's need for touch screens in cars.

You are viewing the 'need' from the end user perspective. That is likely not the driving force that is causing touch screen proliferation.

The likely driving force that is causing it is the manufacturers BOM costs.

A single, rectangular, touch screen, that can swallow up a bunch of different knobs and buttons and basic displays, is likely cheaper from an overall BOM perspective than the set of knobs and buttons and basic displays (and supporting material (mechanical and wiring) for those knobs/buttons/displays).

And, if two (or more) different car models use the same size rectangular screen, then one single "screen" can be used for plural models (with only software differences) vs. the need for unique moldings/buttons/etc. when providing the physical knobs and buttons on different models.

The touch screen also enables different "accessory classes" even within the same model by a software change vs. a different piece of hardware. Think basic hot/cold/manual blower speed climate control vs. higher end climate control where you set a specific temp and the system picks heat/cool/blower speed. The same single touch screen can provide both the 'basic' and the 'luxury' interface via a software change vs. two different physical control units inserted into the dash.

View on HN

> The likely driving force that is causing it is the manufacturers BOM costs.

Whilst I think this is part of it's about more than just reducing the BOM. There's a trend to move all interfaces from knobs and switches to touch screens even when money isn't the problem. For example the SpaceX Dragon capsule also takes this approach which you can see from the images below:

https://www.spacex.com/crew-dragon

A heavily vibrating vehicle which needs to be operated quickly and with 100% reliability by a human operator really should not be operated via a proliferation of touch screens and touch buttons as the images above show.

I think current fashions are very much to blame in a lot of cases. A sleek touchscreen operated car in the showroom is a sexier buy than your more traditional button festooned dashboard. Buyer's only get a chance to regret it after they've spent a lot of time with the vehicle.

> A heavily vibrating vehicle which needs to be operated quickly and with 100% reliability by a human operator really should not be operated via a proliferation of touch screens and touch buttons as the images above show.

In fairness to SpaceX, Dragon largely isn't designed to be operated by a human. There's a reason that page calls out stuff like "how warm should the cabin be" as the sort of decisions being made with these touchscreens.

To quote the page linked earlier...

> Crew Dragon will be a fully autonomous spacecraft that can also be monitored & controlled by on board astronauts and SpaceX mission control in Hawthorne, CA.

I would be very surprised if Dragon did not include the 21st century equivalent of the Apollo 8-ball, manual translation controller and stick.

NASA is a very conservative organization and if there is a lesson to be learned from past near disasters, is how limited is the foresight of mission planners and how infinitely adaptable are the human astronauts.

It has a stick and a couple buttons for most of the functionality that can be useful to the pilots in case the craft misbehaves. It seems to have rotary input so it can manage rotation and translation.

A lot of the controls in an Apollo were to be used in normal manual operations that are automated (or have no counterpart) in Dragons.

There are a few physical buttons for the few critical functions you might want to trigger in the boost/reentry phase. The rest is either 1 g or microgravity and a touchscreen should not be a problem.

BTW, not surre if they implemented it in the end, but they planned to handle screen blurring due to vibrations by shifting the screen content based on accelerometer data in real time. A rather clever idea IMHO. :)

> they planned to handle screen blurring due to vibrations by shifting the screen content based on accelerometer data in real time

IIRC, Orion does that by timing the screen backlight.

If I had a choice between Orion, Starliner or Dragon, I'd go with the Dragon. The spacesuits are much nicer and if I have to die, I'd rather go with style.

Perhaps I'm missing something but how could the Dragon capsule automation handle something like the emergency that occurred during the Gemini 8 mission? Without manual control by the astronauts it could have turned into a real disaster.

https://en.wikipedia.org/wiki/Gemini_8#Emergency

Computing and testing has come a long way since the Gemini days. We now have aircraft (particularly military fighters) that are literally uncontrollable without the computers.

Those control systems only work within certain preprogrammed constraints. They are generally incapable of handling unexpected situations outside of the design limits.

Computerized fly by wire systems still depend on a pilot in the loop, especially in the case of a mechanical failure.

So let’s say the airplane gets damaged, and one of the flight controls is no longer available. A legacy airplane would still try to use that surface because it doesn’t know any better. The F-35 digital flight control systems will say, “That surface isn’t doing much for me anymore, so I’m going to have to compensate by using some other things. Maybe I’ll have to move them a little bit more to get the same effect because the pilot still wants to turn left.”

https://www.airspacemag.com/military-aviation/f-35-faces-mos...

Buyer's only get a chance to regret it after they've spent a lot of time with the vehicle.

Several years ago, probably. Now, when their previous car already has a similar touchscreen?

"Oh, I'm sure touchscreens have gotten better"

They likelly had forgotten how it was before

> I think current fashions are very much to blame in a lot of cases. A sleek touchscreen operated car in the showroom is a sexier buy than your more traditional button festooned dashboard. Buyer's only get a chance to regret it after they've spent a lot of time with the vehicle.

Exactly. Touch screens are pure tech fad in about 50% of use cases at least, with the decision to use them not based on any kind of practical considerations.

However, I'm a little skeptical that the average buyer is going to regret using one. A lot of people seem to get hung up on newer tech == better tech, so they never really seem to look at the newer technology with a critical eye. I often prefer older tech in some cases (e.g. buttons and knows), but my fiancee often doesn't get it and advises me to stop fighting it and go with the trend.

I've been driving my most recent car for 3 years now. I want to go back to analog controls. It sucks. I'm used to it but it still sucks.

Touch screens are not as mature as a button or knob. We know that using a small knob to control a radio works, it is tried and true again and again.

But we also know that replacing the steering wheel with a small knob is a bad idea.

This is because this is a mature technology has that has as many bad implications as good ones.

I guess I’m a way tough screens need to be everywhere since that is how the technology matured and we know where they should not be used.

This argument doesn't make a ton of sense to me. It seems like an inconsistent comparison to compare a single touchscreen to many customized knobs/buttons/etc., when you could also compare a single set of knobs/buttons/etc. to different screen sizes across models.

Car companies don't tend to custom-design knobs/buttons/etc. for each model. They tend to reuse the same ones across many models to save on development and manufacturing costs. They also already modularize these knobs and buttons across different "accessory classes". You can probably see this in your own car even, if you see anywhere on your car's center console or dashboard which has plastic caps where it seems like more buttons or knobs belong. Those are likely buttons/knobs for additional functions in optional feature packages.

Also, even for the same screen size across models, you'd still likely need custom bezels since most models are differently sized and thus have different dashboard shapes/sizes within which the screen is mounted.

Especially with how high-resolution touchscreens in cars are becoming in new models, I doubt they're cheaper than a handful of plastic knobs and buttons.

My guess is that it's exactly what they said. It's probably harder from a software perspective to adapt touch-screen-based interfaces like Android Auto and Apple Car Play to physical controls, and those difficulties probably come up over and over as the versions of these interfaces on the phones are constantly being updated, and the car must continue supporting more and more versions at one time.

Especially with how high-resolution touchscreens in cars are becoming in new models, I doubt they're cheaper than a handful of plastic knobs and buttons.

Every modern car is going to have an infotainment system with a screen. A better question would be whether it is cheaper to have a touch version of that screen or to have buttons for the interface. I suspect that touch would be cheaper than buttons.

Yeah, that's a good point. The question here isn't whether or not to have an infotainment system that integrates with Android Auto and Apple Car Play, it's whether to interface with it via a touch screen or physical input (or both).

Especially now the US and Canada requires reversing cameras in all new cars, there has to be a screen in the car; the question then becomes whether it's worthwhile designing a second infotainment system that doesn't require that (somewhat costly) large colour screen for lower end trims in markets where reversing cameras aren't mandatory, and it seems to be that manufacturers think that isn't worth the cost.

I wish they did that in the EU. Reversing cameras are massively useful, especially in vehicles with low visibility such as the Toyota CH-R.

Really? As much as I like the comfort of the rear radars for big vehicles, I’ve always found the camera a bit useless

As somebody who doesn't own a car anymore but somewhat regularly rents very different kinds of cars, I have to say that reverse parking is massively easier with a camera.

Back when I always used to drive the same car I got used to its dimensions of course, but that just doesn't happen when you merely occasionally drive very different cars.

The sensors are nice, but sensors+camera improve on that by an order of magnitude - you do not see behind your car, no matter how much you think you do. Especially small people aka children are completely hidden from the driver's point of view by the rear of the car - which is exactly in the camera's center of field.

I got a shocking example of this recently. I was towing a trailer behind my vehicle for a 14' boat (the trailer was quite a bit longer than that). I was really surprised that I could not see it AT ALL in my rear view mirror when the boat was not on it. Meaning there is at least 16' behind my vehicle where I cannot see something about 2' tall or shorter.

That's what I thought, but I've rented a couple with camears and they are really useful - the overlays which shows where your wheels will go with your current steering position etc.

The last car I rented had a 360 degree camera which showed a top down picture of the car and surroundings!

I LOVE mine... with the guide lines for reversing, I always center into the space as desired... it's easier than pulling in forward even. For times when the neighboring spaces are wonky, it's easier to short shrift the passenger side, aligned with the other car's passenger side, so worst case you have to pull out to let a passenger in.

Also, if every other car parked forward/reverse then the more cramped spaces wouldn't be as difficult. Note, I live in AZ, USA where more than half the cars are full sized or larger vehicles. So parking lots designed for cities with smaller vehicles are particularly difficult.

Wouldn't it be better just to make cars with good visibility?

Yes it would. That's why I'd never buy the Toyota CH-R. We have a Civic with bad enough visibility already, thank you.

I think that's only half true. There are standardized controls but the dash configurations and wiring harnesses are specific to each model and then even for only a few years. So I'm in agreement with the BOM arguement generally. Also it's a marketing selling point. Consumer wants the shiny new thing.

I totally understand the business case for touchscreens in cars. It's all about reducing cost to the business, but this usually comes at the expense of safety, which is a totally unacceptable tradeoff.

That's a rather communist view. The free market allows people to buy cheaper cars at the expense of their safety, what's wrong with that?

Off the top of my head:

* If the design of your car makes you a less safe driver, you are more likely to crash into me.

* The US as a society has committed (surprisingly recently) to providing emergency care to anyone who shows up at a hospital before asking whether they can pay. Car accidents are a sadly common reason for people requiring emergency care. That means that reducing car accidents is a public benefit.

* Modern societies have made the choice in general to require some baseline level of safety from all sorts of products in order to make commerce more efficient. If consumers feel confident that they can (e.g.) buy a random toy off the shelf for their kid without needing to do a bunch of research on whether it contains lead paint or dangerous parts, that reduces friction and increases sales. The same reasoning can apply to car safety features: it's easier to get people to just buy a car if they have a baseline level of confidence that any car they buy is safe to drive.

That's just a few reasons that first came to mind. I've already thought of one or two more that would take a little longer and a little more care to put into words, but I encourage you to think about these questions yourself when they come to mind, too.

Poe's Law claims another victim.

I wouldn't consider 1986 particularly recent, regarding the COBRA/Emergency Medical Treatment and Active Labor Act particularly recent. Considering that civil rights was still a large issue well into the early 80's. Not that there aren't civil rights issues today, it's mostly that the outrage is much larger than any actual issues persisting.

Consumer protection on the other hand, is quite a bit longer in the tooth. Though did really start after the 50's. I know the original erector sets were particularly bad, as were earlier die cast models, etc.

A. Please be sarcasm..

B. That's.. not communism.

C. Especially in this case, the safety issue affects everyone else on the road, too. If someone crashes into you while playing with their Apple Play, congrats on enjoying the unregulated free market.

If your kid takes a ride in a friend's dirt cheap car they bought because they're 17 and they get in a wreck, congrats.

If manufacturers can make and sell unsafe cars, they will, and it will drive up the cost of safe cars.

When products as critical as cars unsafe, it costs everyone, even those who don't buy them.

I do indulge in sarcasm on occasion yes ;)

Phew

Are knobs and plastic moldings that expensive though? I'm not sure this is really a driving factor. Touch screens have marketing appeal (at least from the seller's perspective), I have to think having a shiny touch screen on your new Mazda is simply just trying to get more people to purchase a new car and nothing more.

There are lots of wires and connectors that come with mechanical controls. Everything need to be certified to make sure you don't have a recall of a million or two vehicles because the air-conditioning dials break after 6 months. All that must be crazy expensive compared to solid state electronics.

None of this is untrue with touchscreens. They also have wires and connectors and must be certified.

And every single application loaded and installed, software library, SDK and license has to be rolled up into the BOM for every headunit sold/installed.

It seems obvious to me that one team of software engineers doing testing and diagnostics on a piece of software is way cheaper than training and then paying entire factories to do testing and diagnostics on each button and knob. (Not to mention we're talking about like, volume controls. Not exactly the most complicated and expensive kind of software to engineer.)

Not to mention if you screw up the software an update is generally cheap and easy. If your button is messed up you have to pay for a mechanic to tear apart the entire dash.

Of course you need to get the touch screen reliable. A touch screen is more complex than a button, but if we call it the cost of 10 buttons, a touch screen replaces more than 10 buttons...

For common functions separate buttons are better for UX. However the touch screen is cheaper.

They have 2. One for power and one for data. Usually they're replacing a lot more than 2 buttons/knobs.

> They have 2. One for power and one for data.

I don't think I've ever seen a touch screen with only two wires. For one thing, even power requires two wires, not one. Data for a touch screen typically requires more than one wire as well, especially considering there are two completely distinct types of data a touchscreen must transmit, those being display output and touch input.

Most touch screens I'm aware of have wire connections more like this:

https://learn.adafruit.com/adafruit-2-4-color-tft-touchscree...

> Usually they're replacing a lot more than 2 buttons/knobs.

Not usually. Other than the Tesla, most cars retain almost all of their knobs/buttons in addition to the touch screen. In fact, the linked article in the OP says they're basically replacing the touch screen input with a single volume knob that lets you additional tilt and push it. Other cars, like BMW with the iDrive, usually have a single knob that lets you rotate and scroll through menus and push to select. Granted, it's not a simple knob, since it requires multiple degrees of input, but even a plastic knob that allows rotating, tilting, and pushing is probably cheaper than a touch screen when shared across models and mass produced.

I still think the bigger savings is probably the software updates required to constantly support and make your vehicle's physical input compatible with every new update of Android Auto and Apple Car Play, which the auto-makers have no control over and can only ever play catch-up. At least with a touch screen, they know that's where the Android and Apple engineers spend their own time making compatible.

The other place it probably saves is designing and laying out the interior, since it's easier to design an interior with fewer purpose-focused inputs. I'm not sure though if the touch screen would actually be easier though, since it can't be broken up across different small areas, and the entire thing has to be within both reach and active field of view. Especially the screens on top of the dash, since those often are designed to retract, which means the space they retract into needs to be designed around the HVAC and electrical systems between the firewall and dash.

Either way, both of these savings though would be on the design and engineering side, not the bill of materials.

Most touch screens I'm aware of have wire connections more like this [...]

That link is to a toy touchscreen. Are we counting every strand of metal as its own wire now? I think the power and the data cables are each bundles of lots of such strands, and no car manufacturers worry about the individual strands because each cable plugs into the circuit board as a cohesive unit, like: https://www.alibaba.com/product-detail/Customized-odm-15-6-i...

(First Google result for "touchscreen alibaba", it's not like this is an obscure, rarely used configuration.)

Either way, both of these savings though would be on the design and engineering side, not the bill of materials.

¯\_(ツ)_/¯ Those sound like good points to me, but I don't know anything about this stuff.

I do know about this stuff (currently work for one OEM and have worked for 2 others) and you're exactly right.

Something that hasn't even been mentioned is all the meetings, documentation, coordination between people that work on the components that have to interface with each other. That labor is significantly reduced when you offload most of that to the touch screen because some of that interface can now be done by the same person without any meetings. That's not just cost savings, it's also time savings, which is part of why Tesla was able to bring the model 3 to market so fast (fast at least in the car world).

That said, I much prefer tactile controls. I recently moved from a BMW to a Volvo had the BMW had physical controls for most common functions while the Volvo only has a handful of buttons and one knob for volume. It's a pretty good balance, but I do miss a few buttons and more could be done on the touchscreen to ease that longing (such as customizable shortcuts).

Cadillac is starting to go back to more tactile controls because they've learned how terrible it is without any (my wife has a CTS so I completely understand why they're doing it -- it is awful). I hope Audi changes course too.

For one thing, even power requires two wires, not one

The chasis is your second cable.

You still need a wire from the touchscreen to the chassis.

Have you ever seen behind the dashboard of a modern car? It's a monstrous tangle of wiring, with inch-thick cabling bundles and dozens or hundreds of connectors. A touchscreen-centric interface like the Tesla Model 3 could easily save several kilos of cables and connectors. There's a huge amount of labour involved in installing all that wiring and a huge number of opportunities for failure.

I support Mazda's efforts here, but there are compelling commercial and engineering reasons to go in the other direction.

Do tell us all about the huge labor savings Tesla is running away with[1].

It blows my mind away how you can even begin to compare the predicted reliability of a complex touchscreen interface stuffed with active electronics to dumb buttons and wires with a straight face, let alone what fielded production units only a few years in have clearly demonstrated.

[1] https://www.thedrive.com/tech/27989/teslas-screen-saga-shows...

No, switches (even good ones), are cheaper than touchscreens that can handle automotive heat and vibration. It hardly matters which you use anymore, since pretty much all control heads are going to be networked onto a CAN bus or the like these days. Touchscreens are a bad idea for cars though, and I applaud Mazda for being the first to try to deliberately kick the damn things back to desks, pockets, and walls, where they belong.

I imagine they are that expensive. Even low volume car manufacturers often "borrow" switchgear from mainstream brands to avoid the cost of making custom switches. This is why any reader fortunate enough to afford a modern Aston Martin will find the cabin filled with a lot of Mercedes switchgear, as one example. Go back farther than that and Aston were owned by Ford, which is why mid-2000s Astons are filled with very cheap feeling physical Ford controls from this era. Etc, etc. Parts sharing is critical to affordable production.

Touch screens are staggeringly cheap now, as anyone with a Raspberry Pi and too much free time can attest, and fits virtually any car, as opposed to switch gear design which often requires the rest of the dash to fit with the design of the button or knob too. A screen mounting location is usually just a rectangular hole in a dash.

> Are knobs and plastic moldings that expensive though?

If you consider the entire cost, yes. Every BOM item has a huge cost associated with it, and it's price only a small fraction.

Consider that for every item you need to:

* secure supply for the next 10 years or so

* organize purchasing and put it into the supply chain

* design assembly instructions, teach workers how to assemble it

* design and implement testing procedures

* design and implement diagnostics to figure out which specific component has failed

* maintain it and keep it available for your service network for the next 10-15 years

* deal with unexpected failure rates and be ready to re-design in case of problems

* also, every knob is actually not one component, because it needs to be connected to something, which implies wires and connectors (the most problematic components in electronics) and multiples all of the above several times.

Look at it this way and suddenly you really want to minimize the number of individual components and replace them all with a single touchscreen. Especially given that you can then deliver crappy software and the market will bear it, because we have been trained to expect and accept crappy software.

But I do agree that the marketing appeal is there, too, although I think some people are waking up.

> Especially given that you can then deliver crappy software and the market will bear it, because we have been trained to expect and accept crappy software.

I'd put it differently; I'd say we've been trained to expect anything with (user-visible) software to be crappy, and to accept it because we have no alternative.

That's especially true for automotive SW and interfaces, which raise dreck to a whole 'nother level. For a whole lot of reasons, I don't expect to ever buy a car much newer than about 2010 for the rest of my life. Avoiding (ob)noxious systems and interfaces is a large part of that reason.

I’m with you. Im not that old but new cars really have no appeal to me. Will be interesting to see how it goes hanging onto a 2010 for 50 years

Me too.

I enjoy driving a manual ("stick") and distrust auto-braking/lane assist/etc. My car has airbags, ABS and traction control so I'm happy with the safety features.

No way will I buy a car where essential functions are touchscreen operated, or with any phone-home "features".

While I tend to agree that the past decade has seen some arguably unnecessary increases in complexity, the latest safety technology is significantly more advanced than even 10 years ago.

Your list includes "Air Bags, ABS and Traction Control" - this list was fine in the early 90s, but you've missed one of the most life saving innovations in recent modern car history - Electronic/Dynamic Stability Control.

Multiple studies have shown ESC has potential to reduce fatal accidents by a third - this is largely why it was made mandatory on all cars sold in the EU and US since 2011/12. If you own a powerful vehicle, especially a rear wheel drive one, the kind of driver input mess (accidental or intentional) ESC systems can clean up that otherwise has you spinning off the road is pretty amazing.

It has "ESP", I assume that's the same as ESC?

Maybe a new Mazda will be up your alley!

Mazda just moved to the top of my "I would buy an electric car from this manufacturer if I could" list, but they don't seem to be in a hurry to release a product.

They may be cheap to create, but the cost of upgrading is incredibly expensive.

It's impossible (without great expense) to add a new physical button to a car as an upgrade. Very easy to push a software update that updates a touch display.

> Are knobs and plastic moldings that expensive though?

Yeah, the molding for any typical single component is tens of thousands of dollars per iteration. A car has a lot of those components, so the fewer the better.

It's not just the up front cost. They can roll out updates and patches to change the functionality too.

Maybe throw in some cheeky DLC to unlock certain features too ;)

I'd like to see a compromise on this. If there's a passenger, they should be able to use a touchscreen. For someone not used to the car, a scroll wheel is going to feel very weird. And obviously safety isn't much of an issue if the passenger is controlling it.

Cars already know whether a passenger is present of course.

Touch screens are awful for operating while you focus elsewhere. I miss touch typing and I sure like adjusting car controls without looking away. What's that, trust driverless cars you say? Never, and the basilisk be damned.

Not to mention that touchscreens rated for passenger vehicles and the extreme environmental changes their interiors go through are not cheap. Look at how many screens Tesla has had to replace in their cars.

It's not that hard to make programmable / re-assignable physical dials and switches to interface in a tactile way with changing software applications.

my old mustang had 10 old fashioned buttons under the screen that i could use to dial a phone number. new jeep has a rectangular thingy that could swallow up everything, but somehow it didn't swallow 10 buttons, so i have to pull up the phone to dial a number while driving.

Reply on news.ycombinator.com