Hacker News Reader

Comment by kkarpkkarp | Hacker News Reader

> Good. There is a reason you don't see apple like designed plane cockpit.

But Mazda's cockpit isn't plane cockpit either. Please go for a test drive with new Mazda 3 and you will be surpised how stupid is this decision.

If you want to volume up, change a song, yes, it is easy.

But if you want for example turn on GPS navigation, you need to use this wheel near the gearbox to: - leave a radio interface (you must look at screen and turn the wheel at the same time) - find where is a navi (by turning the wheel and pressing it, still looking at the screen) - choose, again with the wheel, to start typing your destination - and what most: again with wheel (turning it and pressing) type the address simultaneously looking on an onscreen keyboard.

All that while not looking on the road and holding your steering wheel with one hand and another on navigation wheel on central canal.

How come this is safer by tapping on the screen?!

View on HN

You shouldn't be doing anything like that while driving. Also if it's physical in any way you can build accurate muscle memory and know without looking that you have gone to the right spot. This is almost never the case with touch screens interfaces.

There is no safe way to interface with a screen while driving. You are taking your eyes off the road, and your focus goes with it necessarily. The safest option is no screens. It's the only reason I would never get a Tesla. If it's mission critical it needs to be physical.

It's different in a plane, because you can safely take your eyes off the sky. In fact flying by instruments only is routinely the only choice; night time, weather. So a glass cockpit/touch screen controls make sense there. But even then the key controls are still physical.

Just don't use navigation guy, memorize the paper map! /s

Set your destination before you set off? Or if you really need to do it after that point just pull over to the side of the road for a minute.

Pull over to adjust the map. Does anyone really believe that adjusting GPS while in a moving vehicle is safe?

This gave me pause for thought. It really toes the line of socially acceptable behaviour for a driver, as evidenced by how common it is to do so. How many times have you been a passenger in a car and the driver has pulled over to adjust GPS?

I don't think I have particularly dangerous friends and family, but they rarely do.

> How many times have you been a passenger in a car and the driver has pulled over to adjust GPS?

When I'm in the front passenger seat, dealing with the GPS and the A/C becomes my responsibility.

I’ve never been in a moving car where the driver did anything with the gps more than listening or asking a passenger to adjust it.

Or, you know, stop to re-adjust your course without endangering everyone sharing the road with you.

Your sarcasm looks a lot less smart once you examine the actual options.

Ok Google directions home.

If that's what it takes, then so be it. The yearly road toll is embarassing. It wouldn't bother me one bit if all device input turned off as soon as the car started moving. Passengers can play Eye Spy to pass the time.

navigation was just an example. Angdthere is no place here for muscle memory. Do you want to switch radio station? (I hope I can do such thing during driving?). Look at screen where is focus now and rotate the wheel by some angle which depends on what you see on screen. Press the wheel, rotate the wheel again to select another radio station, press it again.

Things like that. Maybe after few years you will memorize parts of this.

I am not telling resignation from touchscreens is bad. Just Mazda didn't offer anything useful instead of it.

Well I do agree, it's definitely not a great solution either. I am not sure there is one that involves the screen at all.

Media buttons on the steering wheel were peak car usability. What is there really to be looking at other than the road while you drive? A map at best, but we can do better there too. A HUD of upcoming turns projected onto the glass for example. We had HUDs in cars in 1989 (Nissan Silvia) and some modern cars are bringing it back. I think that's the perfect blend of usability without compromising safety.

It doesn’t matter what you “shouldn’t” do when it comes to mass-market consumer devices. It only matters what people will do, especially when you share a road with them. And when it comes to vehicle interfaces, you should assume 80% of drivers will adjust them during a drive, no matter how inconvenient and dangerous is. Any system that doesn’t recognize this reality — and try to minimize the harm if it — is just going to get people killed.

How you do design for this consumer who is controlling while driving without encouraging the activity by making it easier?

Remove the ability to control the screen while driving.

Well that's not really designing for those users then. Your original comment said there's no way to interface with a screen safely while driving, but the other user said designers need to design for those dangerous users. I was wondering how you could do that in a safe manner (I don't think you can personally, and agree with the no-control-while-moving idea).

> but the other user said designers need to design for those dangerous users.

The other user said you had to design for what people will do; disabling all access and removing the choice to interact is exactly doing that. Disabling hands-on interaction above a certain speed but allowing voice controls might be another option.

The best I can come up with is transparently overlaying small information cues on the glass of the windshield, and using sound cues. I think in general the systems available to a cars HUD should be pretty limited and the controls should be on the wheel.

I totally agree, which is why I think any screen is a bad idea. That ship has sailed though and we have millions of cars with screens, so it's still worthwhile to keep repeating that it's ridiculous to use the screen in motion, even if that's what it was designed for.

They can make most things non responsive while the car is in motion. This would inconvenience the shotgun passenger but typically the driver calls the shots on the radio (in-car entertainment system) anyhow.

Unless a law is passed restricting manufacturers from allowing responsive interfaces while driving then any disabling of them will be a significant market disadvantage. No sane car company would do it.

At least ford and Toyota already do to some extent.

Some manufacturers have had restricted interfaces in cars for a while now. On Ford's SYNC system I have not been able to type in an address if the car is moving. Android Auto similarly limits inputs to voice inputs on any kind of text entry while the car is in motion.

> You shouldn't be doing anything like that while driving.

What does that have to do with anything? The touchscreen controls for most of that aren't even enabled when driving.

But on the Mazda you can't even do these things with the touchscreen while NOT driving.

This entire article, arugment and comment thread is about how Mazda is removing touch screens to improve saftey while driving. I agree the Mazda solution isn't much better though.

The person that I quoted was responding to someone who talked about keying in directions and the limitations that Mazda's limitation will cause for that. That functionality is not available while driving even with a touchscreen.

To me, its a situation of throwing out the good with the bad.

I understand the thought, but in the real world this is counterproductive and dangerous.

Some interfaces it’s pretty clear that they are best served by physical buttons, toggles, switched, etc. Blinkers, gear selector, volume, windows, for example.

Some interfaces it’s equally clear that they are best served by touch. Text input through an on screen keyboard obviously beats a scroll wheel to select individual letters. Likewise panning and zooming a map.

Now, whether some of these interfaces are being operated while parked or driving is actually a completely separate matter. As you note, any touch feature can be dynamically locked out from operating while in drive. But to answer that question you can’t take the position of a nanny chastising a child. You have to take a holistic view of what actual users in the real world will do based on your decisions.

So, navigation is a great example. People need to navigate while driving. Sometimes that means interacting with the navigation interface while under way, whether it’s due to updating a route, checking for traffic, zooming in and out. As much as you might like to, car designers don’t get to choose whether drivers are going to do this activity while driving. They can provide an interface that is as safe and convenient as possible, or drivers will just use their phones instead.

Every car nav system I ever used, including fairly recent Audi and Mercedes ones, that depended on physical buttons for input was universally terrible. So bad that it wasn’t worth using and so the phone was used instead.

Compared to the Tesla Model 3 nav system which is fully touch, highly responsive, and beautifully rendered. It is a total joy to use and increases overall safety through large touch affordances which make sense for the driving environment. And it’s so easy to activate I almost always have my destination plugged in, which means the car is showing me upcoming traffic and helping me avoid accidents through alternate routes, which is also increasing safety and decreasing overall traffic congestion.

I think we have pretty similar views while talking about the tech and usability itself. It is just that I fundamentally disagree that you need to touch or look at anything while driving.

You say people need to interact with navigation while driving, which is simply not true. People want to interact with navigation while driving. They could cope without it. We didn't even have navigation all that long ago, to say we need it is a stretch let alone to touch it while driving. If it's causing accidents, and it is, it should be considered dangerous. Perhaps the Model 3 is better at it, but it is still a danger to look away from the road even for a second or two.

That means screens are out, and we need better ways to display and interact with information while driving. HUDs, sounds or feel is all I will agree is safe.

I said people need to navigate while driving, which is almost a tautology. That is to say, while I’m driving I need to know where I’m going. Sometimes I already know from memory, sometimes I’m consulting an external source.

Whether that external source is a sticky note on the steering wheel, a phone, audible prompts, or a big touchscreen in the center of the vehicle (TM3), the driver is going to decide which to use based on the options available to them at the time.

What the driver is not going to do is pull over at an exit to stop the car and check if that was the exit they were supposed to take, and then get back on the highway and keep going.

The absolutist stance does not actually increase safety. In the absolutist stance, well people don’t need to drive at all. To design safe systems in the real world that save lives you need to provide an alternative that is both safer and as easy to use. Because a safer alternative that drivers don’t use by choice because it’s too annoying or inconvenient doesn’t actually improve overall fleet safety.

Obviously when you operate a vehicle you don’t always just look straight out in front of you. You check your mirrors. You check your blind spots. You maintain situational awareness and do that safely by planning where you focus your attention from one moment to the next. Screens are obviously not “out”, nor is it correct to say that looking at a screen while driving is unsafe. Instruments, whether for navigation or simply indicating current vehicle status such as fuel, speed, lights, etc. are an integral part of driving and digital instruments are not necessarily less safe than analog, and can potentially be more safe.

And so returning to my original point, in my personal experience, the large touchscreen in the TM3 is ideally suited for safe navigation, and improves safety when navigating particularly on an unknown route, by presenting crucial information in extremely visible and usable form, and allowing me to configure and interact with the system naturally and efficiently while I proceed along my route.

> What the driver is not going to do is pull over at an exit to stop the car and check if that was the exit they were supposed to take, and then get back on the highway and keep going.

Funny you should say that, its exactly what people did not too long ago.

You'd look up where you were going in a key map or more broadly a road atlas, write down your instructions (and keep it open to the final destination in case of road closures, etc)

I really don't see the appeal of getting in a car and beginning to drive without knowing where you're going. "I'll look it up on the road" just seems like an intersection of anxiety (deliberately increasing your chance of crashing), annoyance (deliberately making your navigational search harder to perform), and apprehension (increasing the chance you'll miss a crucial turn early in the trip due to lack of navigation).

It's like choosing to juggle while taking a test.

The hope is that they have entered their navigation correctly before they left and from the time they leave 'til the time they arrive they don't need to look at the screen at all, it's all voice and smaller visual cues. I guess I just don't feel like it's a stretch to have the responsibility to figure these things out before embarking, instead of doing it on the fly which is more dangerous.

The absolutist stance could absolutely increase safety. If we ditched all screens cars could only be safer. Less convenient, but safer. I'm not saying ditch navigation or music though, but ditching off to the side navigation/media screens could only be a plus to safety. Implement navigation and media in a way that's more sympathetic to the need of remaining attentive. Any lapse in attention we can get rid of is a good one.

You haven't argued for how having screens could be safer than not having screens, only for how to improve the safety and convenience of screens we have. Your advice is probably going to be more useful, because in all likelihood we'll keep screens. But I wouldn't care if it took someone 45 minutes longer to get somewhere because they got lost, and had to pull over and ask a man sitting on a bench for directions. Driving is a privilege and we shouldn't trade safety for convenience. I think media/navigation screens could be replaced by better solutions that don't require a screen. Who says you need a full visual map at all? Who says you need to type in the address? These are problems that could be addressed, and I feel as though the screen is the lowest effort solution not the right one.

> We didn't even have navigation all that long ago

Yes we did, even if it wasn't an electronic system built into the car, and it wasn't unheard of for people to pull out physical maps and reference them while driving, or pull over and stop in unsafe locations to reference maps. N Electronic navigation systems are what make people less likely to do those things. There's a difference between introducing a digital tool and introducing the activity it supports.

I meant digital navigation, but fair point. Trying to interact with a digital map while underway is still dangerous, like trying to use a physical map was. Listening to a GPS is great, and I'm not advocating against digital navigation completely, just devices that require you to shift your attention away from driving.

You seem to be equating touchscreens with safety and quality. These do not go hand in hand - the underlying system can be absolute crap with a touchscreen (remember old dedicated GPS units?), or they could be Tesla-equivalent but with a weird physical button input system.

> Sometimes that means interacting with the navigation interface while under way, whether it’s due to updating a route, checking for traffic, zooming in and out.

You shouldn't be fiddling with the navigation system while driving - you should do that while you're still in park. Even when I look down at my phone to see what lane I need to be in for a second, it's amazing how much distance passes when going at highway speeds. Unless you are using autopilot, it's dangerous to mess around with the map, and it's not that hard to pull over to do it, either.

Oh I definitely understand that the vast majority of touch screens are utter trash! But that's not to say that all are.

Understanding what lane I need to be in, and what the traffic conditions are up ahead around the bend; these things add to situational awareness that makes driving safer. But I'm just not going to use a shit GPS that I need to program an address by spinning a scroll wheel and individually selecting letters.

If you had a large touchscreen in your car, you wouldn't have to look down at a tiny phone screen to see these things. They would be in your peripheral vision and easily checked without losing track of the road.

Now I regularly move to the right and slow down long before approaching a traffic jam on the highway, versus the typical hard-brake when the traffic comes into view.

At an unfamiliar complex intersection (greater than 4-way), sometimes you can't figure out where the GPS wants you to go. A quick pinch-zoom before the light turns can give you the visual queue you need to avoid having to make a last minute decision as you enter the intersection. That's totally possible on the TM3 because it's instant, smooth, and fluid (and it's really great at automatically zooming back out to the right level as you start driving again).

And to give another anecdote, today on my drive home navigation was set to take me 'Home' (by the way that's two taps on the screen, and takes less than 1 second to activate as soon as you sit down in the car). I was on the highway in stop-and-go traffic with AutoPilot. My wife called and asks me to pickup takeout. I know it's too early to order, and traffic is terrible. I need to know how long to wait to order, so I tap the steering wheel control and use voice control to change the destination. Because it's internet connected, I can just speak the restaurant name and the town. It tells me the traffic-aware arrival ETA so I know when to place the order, and adjusts the route (taking an earlier exit). Unfortunately there's no single tap on screen to make the call - POIs don't have the phone numbers listed (yet)!

So I find that there are plenty of perfectly appropriate reasons to adjust navigation while driving, I just think very few people have experienced how a proper large touch-screen implementation can supplement and enhance your situational awareness and allow you to complete tasks safely and efficiently while underway. The is strictly better than the real-world alternative which is for people to try to complete these tasks using a phone-in-hand while they are also attempting to drive.

>At an unfamiliar complex intersection (greater than 4-way), sometimes you can't figure out where the GPS wants you to go. A quick pinch-zoom before the light turns can give you the visual queue you need to avoid having to make a last minute decision as you enter the intersection.

Seems like a usability problem. Google Maps displays what lanes go in what directions so you don't have to zoom in on the map and will speak it to you so you don't even have to look at your phone/Android Auto/etc.

It does show an icon representation of the intersection and which lane to be in. I just prefer zooming in, sometimes Boston intersections are quite complex.

You can tap on the bubble in the search results to display information about the place and it will have a call button there. Don't remember if it still has that bubble once you start navigating.

It is also supposed to give you POI info including the phone and working hours if you tap on any POI bubble on the screen, but it doesn't work for some reason.

Check out the control panel on a BMW X3. There are dedicated buttons for navigation, media, menu etc next to the big scrollwheel. Navigation even has two buttons. One for showing the map, one for launching a new location.

The screen is touch enabled but I cannot remember the last time I actually touched it.

Mazda has shortcut buttons as well. See: third image in the article. Their UI is very similar to BMW.

Your argument doesn’t make sense to me. Why would you do any of that WHILE DRIVING?? That’s both dangerous and foolish.

As for key presses to navigate a menu you would be surprised how quickly a human memorized exactly how many times to press what buttons to get to where they want to go. In my current car I know that if I press Home, down x3, and OK I will end up in the Media menu. I can do all this without taking mu eyes off the road (or even thinking about it much after a year+ with this vehicle.

Incidentally, that was the aspect of old feature phones I most miss today. I could operate them without looking by muscle memory - and it worked because the UI was fixed, but more importantly, it was soft-real time: the delays were predictable and you could learn them too. Today's touch devices tend to lag completely randomly.

>it was soft-real time: the delays were predictable and you could learn them too.

Wow, well put! This is definitely something that we lost with the current gen of interfaces, and this articulation didn't occur to me.

Losing keypresses or clicks/touches at random, unpredictable freezes and/or delays on interactions, all this contributes to a general sense of unreliability and not being able to infer the state of the device. Consistency is actually a much more important metric than just plain latency.

There were green-and-black text UIs that allowed receptionist at hotel check you in blazingly fast, keeping an eye contact with you. No mouse, just keyboard shortcuts.

Yup. These are a source of a frequent disagreement between myself and some UI/UX folks. I believe interfaces like these were/are superior to modern ones, precisely because they allow one to work efficiently.

I think it's an easy cop out to say, "Doing literally anything other than focusing exclusively on the task of driving is not worth talking about."

The reality is, people will be adjusting their radios while they drive. They will be looking for open restaurants, changing their destination, and a number of other things while driving, and pretending like they're not worth accommodating ignores some very fundamental principles on the topic of economics.

People do, because they can. It doesn't matter whether they should or not.

When you design something, you don't design for ideal use, you design to eliminate risky misuse. This is a core axiom of user interface design. It's not good enough to just slap a warning disclaimer on it, as the designer, you're still liable.

I would have my engineering licence revoked if I ignored possible avenues of misuse, could design them out but didn't, and then ended up in court because they misused it and someone died. This isn't a hypothetical, this is how it works in the non-software world where engineering decisions can routinely kill people.

Maybe, the point is not to make it easy. Maybe the point is to make you think "Hey, there's no safe way for me to navigate through 10 different screens to get to my GPS while I'm driving. Maybe I should pull over and do that instead?"

Most touchscreen varients are actually worse than this, and it's nessecary for their displays to be within arm's reach, increasing refocus time too.

I think this just proves there are no safe manual interfaces for setting a navigation destination. Drivers should use a voice interface or stop the vehicle in a safe location.

Doing that with a touchscreen or physical controls while driving would be illegal in most jurisdictions, no?

> But if you want for example turn on GPS navigation, you need to use this wheel near the gearbox to: - leave a radio interface (you must look at screen and turn the wheel at the same time)

Did you miss the part where they said they will be projecting an HUD onto the windshield so drivers won't have to look down? Your comment is arguing against a solution Mazda did not propose.

Curiously, most of the people who are defending touchscreens in this thread are not addressing the 2 safety issues brought up - inadvertent torque applied by drivers and inattention to the road ahead.

I don't think that solves the issue.

whether you're looking down or forward, you're spending your attention looking at the HUD, not the road or any other sensor.

Reading the article is always strongly recommended. The research conclusions are in there.

Did they change this on the new Mazda3? I had a CX-5 and you have buttons that took you to those places (audio, gps, and two more I don't remember) around the wheel thingy.

I am thinking on their new (soon) CX-30 as a possible new car as I loved the CX-5 and was told it would be mostly be based on the Mazda3.

No, they didn't change it. Buttons are slightly re-arranged, but they're still there. See the 3rd pic in the article.

If you want GPS navigation, you press the "road" button:

https://images.hgmsites.net/lrg/2019-mazda-mazda3_100689497_...

You don't even need to look down to do it, because after 10 minutes of driving in your car, you've memorized the buttons.

To be fair, navigating with carplay is also highly distracting because the voice controls are so basic and it forces you to pull over and deal with one of the two touchscreens to deal with route changes, adding a stop, evaluating express lanes, dismissing modals, looking at upcoming directions, etc etc. Navigation is highly distracting period.

The proper way to do this is by voice input.

Voice input in cars is the worst idea I've ever seen. I've had a brand new Mercedes with the MBUX interface and extensively used Android Auto, and the voice commands just don't work. Half the time I'm sitting there repeating myself like an absolute idiot because the car completely misunderstood what I mean(yes car, "Navigate to Cramlington" totally means "Navigate to Boston, US"). I have so many horror stories of android Auto getting stuff wrong it's almost funny. And then it's just made worse when you are bilingual - voice interfaces suuuuck at working with two languages at once. Try saying "call Szczepańska" - guaranteed it won't work. Then try getting it to read a message in any language that isn't English - it's just garbled nonsense. My favourite one was when my wife sent me a link to something and Android Auto started reading the link out ....as H T T P colon slash slash W W W dot......what sort of madness is this. Who would ever want this crap in their car is beyond me.

It can be done properly. The BMW Mini has a pretty good voice input for the navigation, it relies on spelling out the names of cities and streets, there is a button based back-up but I find that the voice recognition works more than well enough for almost all destinations.

Voice output that spells is silly (as is Google's navigation, which overwhelms you with information you do not need).

Just because the technology sucks right now does not mean it's a bad idea. I can EASILY see voice commands in cars being commonplace and far superior to anything that exists right now.

So why not wait until it works instead of pretending? They sell it as if it is perfect right now, the future has come, except it's not.

Guess you could call me trilingual (swedish, english and spanish) and I'm having no issues with Android Auto understanding destinations even when my spanish/english is broken as hell, with music playing in the background or engine noises coming from outside.

Maybe voice commands are just not ready for polish (guessing Szczepańska is polish) but seems to work fine for at least english and spanish.

I just drove 14 hours this weekend and I can confirm the voice controls are almost entirely useless. If you’re lucky it interprets you correctly and you have a menu to sift through on a screen to pick the right option. If you’re unlucky it just does something without any confirmation. The only success I’ve really had is “direct me to <large city here>” and “direct me home”. All other efforts lead me to swear and pull over.

Granted, I was a child map reader and navigator for my parents, this isn’t easy for humans either.

Which doesn’t really work when you have screaming children (or adults) in the car.

Maybe that's a particular Swiss problem, but I suspect it's actually the case in many regions (as strong local accents are a thing): I've never seen voice input that works for place names. My current car has pretty recent technology and its voice input is great for calling someone from address book, it can even understand Japanese names. But Swiss place names? Forget it, it's impossible for the standard German interface to pick that stuff up correctly, and I'm trying to speak in the most accent-free way I can manage.

Pronunciation of the names is the problem, spelling them should resolve that issue completely.

do these systems automatically pick up if I spell rather than pronounce, or do they need a special mode? I've never tried that...

I never tried the pronunciation, I'll give it a shot and see what it does. At a guess the results will be terrible :)

Reply on news.ycombinator.com