The designer gave a lot of thought to the problems with current touchscreen UI in cars, and kudos to him for doing so.

However, it's not the solution yet, in my opinion.

First, all the methods of interaction are invisible. There are no cues to tell the user what two fingers vs. four fingers will do, or what the difference is between a close-fingered gesture and a wide gesture. At best, you'd have to standardize the meanings across all cars (something that rarely happens in autos, even today there's no standard for where the wiper controls go, for instance), and even then, people would have to memorize those invisible gestures before they could use their car.

Secondly, accidental inputs would happen with some frequency. People have a wide range of hand sizes, motor control skills, even number of fingers. Suppose I've got two of my fingers in a cast? I can use an existing car's dashboard just fine with that temporary impairment, but not this.

Finally, it's addressing a problem that already has a solution—physical, dedicated input devices. Humans have spatial memory, we learn where things are, and learn to reach for them without looking or even thinking. Your muscle memory tells you how to reach the wiper stalk or the gear shift, which is why it's sometimes disconcerting to get into a car with the gears on the steering column if you normally shift in the center console. We need in-car systems that let the physical feedback, affordances, and reliable location of buttons, knobs, and switches interact with display systems that are designed for the attention a driver requires. You should be able to keep your eyes on the road and still adjust the air conditioning—something the designer in this video recognizes—but that shouldn't require learning invisible gestures that are prone to user error.

This is way better than what they're building today, but honestly, I think a panel of PHYSICAL buttons, knobs and switches is even better.

The iPhone-style interaction -- a flat screen of glass where you manipulate "controls" with touch sensors -- has one major advantage: it allows an INFINITE number of different controls. (Every app is its own set of controls.) But to do this it relies heavily on continual interactive feedback with your vision.

When driving, you can't do that continual feedback. Perhaps someone clever can figure out how to do interactive feedback with audio (you can use that sense somewhat without taking your attention from the road). But otherwise, you can really only manage as many motions as your fingers can "memorize" (muscle memory mostly). If you're doing that, why not stick with switches and knobs, where your tactile sense can help out.

Everything he shows here is accomplished already by a fairly nice set of buttons on my steering wheel. Each button has a unique shape, so finding them without ever looking down is very easy, and my hands never leave the wheel.

What I find hard about in-car UIs are things like entering and editing navigation directions or selecting music to listen to given a 32GB USB stick full of songs.

There's always some horribly designed onscreen keyboard or "iDrive" type knob controller, both of which are quite horrible experiences. To make matters worse, the UI is laggy and unresponsive, and in an effort to make things "easier" onscreen controls frequently dim & disable when not actionable or when the car is moving vs. stationary.

Agh! Do we never learn?!

http://worrydream.com/ABriefRantOnTheFutureOfInteractionDesi...

His examples are unfortunately pretty terrible. Volume control and air-con are all settings that you frequently adjust. These are not the controls that should be hidden in a touchscreen.

It's the dynamic, changing display of touchscreens, where you can navigate through a menu (or other structure/organisation) of settings, are where the real benefits lie. If, like in the video, you don't make use of the screen, you're left with a tablet-sized void on your dashboard. Most people agree that the existing buttons and dials are far better, easier and intuitive to use, so your volume & fan speeds are going to be best served by physical controls.

The problem is with the less-used controls. Cars have lots of settings and if you add a button or dial for each of them, you'll end up with something looking like an audio mixing desk. This is where touchscreens and displays can improve the UI. A well thought-out system should let the driver find any setting they need and let them adjust it. Leave the common controls as-is, and work on how best to present the others on the screen.

Interesting concept but I am concerned about using multi-touch to select devices. First off it's hard to remember, and secondly, if I am missing a finger, I can't set the air-flow rate in my car? Also the differentiation between close and spread fingers is going to be tricky for a lot of people to remember.

Edit: Another thought... the UI for many funtions needs to provide visual feed back... say for what system I am selecting or what station I am tuning. The text is quite small which would cause even more of a distraction to try to read. Yes I can begin the selection process without a visual cue, but I still need to look to see what I'm selecting. Not sure I see a large improvement in driver distraction there.

Slick for users like us? Definitely. Intuitive for the average Joe out there? I don't think it would fly. Too much to remember and too much dexterity required. My Mom could never use this. At best I see this as an alternate UI the driver could select over something more conventional when they feel comfortable with it.

Kudos for thinking outside the box though, there are definite UI gems here that could be leveraged.

Touchscreens are not only dangerous to drivers, but fairly unusable unless you can make judgements based on touch alone. I've used this[0] system before and it satisfies both of those requirements. Can it be optimized further? Possibly. But I don't beleive touch screen is the future (until we get driverless).

[0]- http://www.bmw.com/com/en/newvehicles/x/x5/2013/showroom/dri...

Others have said it in softer terms, but IMO this is a horrifically bad design.

The fact it's not implemented in a car is why it seems like good design - because sitting focusing on your iPad with no other way more important task to deal with means you can 'figure it out'.

How do I tell a casual driver of my car how to adjust the A/C? Or the radio and GPS? How do I tell - at a glance - which control I need to hit? I can't even hit a control - I have to somehow apply steady constant pressure to the screen to execute a gesture - so that's completely incompatible with anyone who likes their seat a little further back.

The use of any amount of press and hold gestures is completely unacceptable, yet that's the first thing we see on the screen.

Until touch screen controls have haptic feedback, I'll much prefer physical controls.

Same reason why my phone has a physical keyboard.

This is the problem with "modern" society / "invention". Touchscreen is not an improvement. It is in fact a downgrade. It adds nothing to the driving experience, and makes the device harder to use.

To quote Antoine de Saint Exupéry, "It seems that perfection is reached not when there is nothing left to add, but when there is nothing left to take away"

Modern "convenience" is more like modern pain in the ass.

My car is a 1991 and has manual locks, manual windows, manual everything.

Beyond the radio the only buttons in the entire vehicle are the headlights. There is a button for the lights, and one for the brights. Pressing the one for the brights is rigged in a way that physically enables the bright button if it is not already depressed. High tech.

The heat controls consist of two vacuum powered left to right sliders.

I find myself want for nothing. Nothing. So simple. Nothing to break, nothing to worry about. It gets me from home to work, and back, reliably. It does what a car should.

The designer evidently hasn't looked into any of the multitouch research. This interface would actually be far harder to use, learn, and get accustomed to than existing interfaces. It's unintuitive. There's no reason why 2 fingers should correspond to volume. With loose coupling like that, and no standards committee, hopping from one car to another will be extremely annoying.

Interfaces need to have strong couplings between action and zero overloading. Two finger ubiquitously means "scroll now" on trackpad. But even on an iphone, multitouch is seldom used for this reason. A trackpad has no images to display, so it is forced to overload gestures with multitouch. Phone don't have this limitation, and thus, intelligently avoid the use of multitouch.

As another anecdote, mobile games that rely on multitouch tend to not do so well.

This is hands down far, far better than what's out there already. A lot of thought went into this and it shows.

Could there be a few options to enter other modes, though? What I mean is that right now there are only 8 adjustments possible. That's better than the 4 I thought it was limited to, but still nowhere near enough for modern cars.

I would estimate that on a modern car you need to be able to adjust something like 100 parameters. Only a few of them frequently, but definitely more than 8 in total.

One of the thoughts I had was that using a (say 90 degree?) twist motion along with a number of fingers could bring up sub-menus. That way if you want to adjust other radio parameters like fade, balance, equalizer, etc you would put two fingers on (for volume) and twist.

Maybe you'd arrange the different controls in a loop so that every twist indexes to the next and there's a ~10 second timeout before you head back to the default. So if you want to adjust all the "volume" parameters you put two fingers and twist right once for balance and adjust, twist right again to get fade, twist right again to get highs, etc. Make a mistake? Twist left to go back and re-adjust.

If you want to get out of one of the sub-menus you're in prior to the ~10 second time-out for default, another gesture? Maybe a double tap? I don't know on that one, really.

Car manufacturers need to give up on the idea that one screen should control everything in the car. A touch screen just isn't suited for every task. Knobs are the best volume control and this is unlikely to change. They are also a great way to control air flow volume and temperature. While a touch screen could be great for selecting music to play, the other functions of the dashboard shouldn't be shoehorned into a touchscreen.

I'm not sure that the New Car UI is better than what car manufacturers have already done. None of the functions of this interface can be easily discovered. You would need to watch a training video to learn how to use one of these. It would be too easy to forget which gestures correlate to each function.

There's so much to say about this video, but I'll summarize it by quoting one of my favourite articles: "Solving the wrong problem". [1]

(By the way, itt's funny how nowadays, hardware UI is immediately associated with "touch controls".)

[1] http://prog21.dadgum.com/130.html

Perhaps I'm just not envisioning it correctly but this seems to replace looking at the screen to touch a button with looking at the screen to figure out how far apart my fingers are, or how many fingers have touched the screen after hitting a bump. I'm not yet convinced it's a great improvement, though an interesting idea. I'd like to see a demo in a more realistic environment, i.e. in a car on a non-perfect road rather than on a table.

Neat, but I'd imagine that this represents a steep learning curve for many. Say that I borrow a friend's car that has this. How do I figure out how to change the temperature? Touching five fingers close together is easy but not intuitive.

I personally feel one should be able to locate and modify any automobile control without having to look away from the road.

I find it interesting that the Boeing 787 uses five 15 inch displays, but for information display only. Input is still done with buttons and knobs.

[1] http://www.boeing.com/commercial/aeromagazine/articles/2012_...

This may be quite hard to learn for a while. Especially the less often used options. I'd be interested to give it a go though.

For now the best car interface I've used was voice recognition. It usually works great, because it's not a general purpose text dictation. It only needs to recognise less than 100 words and it's not that hard to do. For example I'm not a native English speaker, and I can't get google voice input to capture most of the sentences correctly - it's simply not usable for me at the moment. But I'm quite happy to use voice commands in a Prius - I don't think I've ever used more than 4 or 5 of them, always perfectly recognised.

Creative ideas. However, discoverability is poor. Can you imagine getting a rental car or borrowing a friend's car and getting this to work?

Also, this uses a whole iPad screen to control 8 settings-- eight knobs would take less space and maintain the benefits of a physical interface.

Personally I think commonly-used controls like volume, temperature, fan speed, and some music controls should have physical dials. You can then find them by feel, they're reliable and modeless, and they're much easier to understand. Then, leave the touch screen for navigation and more advanced features.

If you're driving, you don't want to be looking at a touch-screen and you certainly don't want to be simultaneously trying to recall how many fingers turns the radio off! Physical buttons utilize spatial memory and offer haptic feedback, allowing users to keep their eyes on the road most of the time. Removing physical knobs, switches, etc. and replicating their function with a center-console touch-screens is a horrible thing to do to drivers, no matter how good the interface is. For this reason, the center-console touch-screen is unlikely to be the center of car-interfaces for long.

Okay, so what's better than a center-console touch-screen for user feedback? A transparent display either in the windshield or just behind it. Yes, a fighter-jet style HUD. The user could be twiddling buttons on the dash or smearing his/her greasy-finger across the inner surface of his windshield (only for infrequently used functions). The important thing is that user feedback comes from the same place that a driver's eyes ought to be directed when driving!

P.S. As a resident of a frigid country, let me just add that the less I can control with a pair of mitts on the less likely I am to buy your car.

Here's a brilliant completely new idea: what if you had big simple physical buttons and knobs?

Oh wait, that's how it actually worked before we've broken it…

I like the design of the UI and think that there are a lot of good ideas here that could be used in other domains. But you started the video saying current designs are filled with intangible buttons and are therefore hard to control and are distracting, and I think there are some fundamental issues with this design that do not solve these problems.

If I'm already familiar with either my current car's touch screen UI or your design, I can do anything pretty quickly. So the real question is what happens when I don't know or remember exactly how to operate it. Say I'm driving on the highway at 60-70 mph, and I want to change my radio input. How many fingers was that? I can either try all the combinations, which would be incredibly distracting. Or I can figure out how to bring up that legend (better remember how though since there's no indication) and stare at the tiny buttons. Now, since only one option is displayed at a time, I need to figure out the spacial relationship between FM and AM. Is AM a swipe up or down from FM? Is there something in between? I need to constantly be checking the dashboard as new selections are made. This sounds incredibly distracting.

And I'm expected to do all of this while driving at 60 mph. Granted, much of the current touch screen UIs in cars suffer the same problems. But you haven't solved those problems in any meaningful way.

> All in all, this interface gives you easy control over 8 different settings. And it does that without you having to take your eyes off the road because you're being distracted trying to hit that one small button on the screen.

I have a hard time believing the vast majority of people would be able to operate this interface without taking their eyes off the road for a significant amount of time.

I have a hard enough time getting multi-finger gestures to actually work on my iPad and that's while it is right in front of me with my full attention. I can't imagine this all happening while trying to drive the car too.

I think we have too much shit distracting us in the car as it is. I am more than happy with physical dials and sliders for A/C, dials and buttons for the audio, and hopefully someone will come up with a descent Siri-like voice command input for messages/voice and GPS.

Asking me to interact with a blank screen and having to look at it to make sure it's doing what I want is asking for trouble. Just the other day a cop was walking the line-up at the traffic lights and booking anyone who was fiddling with their phone.

Voice control or GTFO. It's already dangerous enough changing the radio channel with the physical button I know by feel. Touchscreen with no tactile feedback sounds like disaster.

I am fond of the controls in my car. Pretty much everything is done through a simple knob with a label above it.

On the far left, a knob for lights. Pull out one click to turn on parking lights, all the way out for headlights. Twist left and right to adjust dashboard brightness.

Next to that, fan control. One click out for slow, all the way for fast.

Just to the left of the steering wheel is the heater control. Pull it out for low->high heat on a sliding scale, rotate it to control the defrost.

Immediately to the right of the steering wheel is the cigarette lighter.

To the right of that is the wiper control. One click for slow, all the way out for fast.

Center of the dash is the radio. One knob to turn it on and control volume, one to tune.

You might have guessed that this isn't a modern car. It's a 1962 Studebaker. The controls are simple, clearly labeled, and do everything you need. You don't even have to look down to find the right control, you can just count the knobs from one side to find the right one. You can easily tell the state of the heater or lights because you can see and feel the state of the control. You don't really even need fine motor control except for tuning the radio, and that's pretty forgiving too. Were I designing a car interface, I'd consider the lessons presented there.

I really like this. My only issue is that the prototype ignores single finger, which to me feels like it should be the most important control. The other issue with number of fingers based gestures is, well, people who are missing fingers. I also think using both the X & Y position could be really useful. Why not tune the radio by dragging up and down and adjust the volume by move right and left, or other combined gestures?

I came up with a similar solution while doing research on "eyes free touch controls" for a certain multinational engineering and electronics company a few years ago.

We found that these types of control perform well, because they effectively have an infinite surface area, but they are not very intuitive, especially when the number of fingers is used to distinguish between different controls. We ended up settling for slightly less "mystery meat" controls similar to Apple's UIDatePicker, which offer a good compromise between being immediately clear and offering a large initial touch surface (which is all that matters, because as soon as you are scrolling the control, your finger is tracked even if it leaves the component).

Touch screen interfaces still have a long way to go before I consider them safe for use in a car though. I did some testing with an eye tracker as well and found that people still glanced at their hands a lot (even after adding artificial tactile feedback). I wasn't able to account for prolonged exposure / experience in my experiments, so maybe it gets better after training.

Now mount the iPad in your car and try using it while going at 50 mpg over even slightly bumpy road. Physical controls give you tactile feedback, but more important, they give you something solid to hang on to while using the controls. Touch screen is good for GPS or settings screens. But often used knobs like heat settings, lights, and such, should be physical.

I like this attempt to rethink the UI for vehicles and I think you are on to something. The absence of any cues I guess is still a problem and the finger-distance based approach could be a problem for some.

How about selecting what you want to control (Temp, Channel, Air etc) on the first pass followed by adjustment of the selected control on the second?

Tangent - because I believe I'm a minority here on HN - a hobbyist mechanic.

Right now one of the worst things to see are these all in one NAV/radio/media/HVAC devices. They are basically impossible to troubleshoot without specialized equipment. Only dealerships have these - the private garages don't see enough volume of particular makes and models of cars to warrant a purchase.

As this becomes more and more common I may even double-down on learning how to fix these things, because it will very quickly become an essential skill in the not too distant future.

Think briefly about the instrument cluster that was in the car you drove up until about 2010. Then do a quick skim of various manufacturer sites and examine the kind of clusters that are going into even new Dodges and Fords - not to mention luxury cars.

Can you imagine having to go to a junkyard to pull out an instrument cluster for a 2011 Ford Festiva in the year 2020? What a nightmare.

The most common buttons are already on the steering wheel, not the touch screen, so I really see limited functions that this can manage. The big problem I see is that there is a lot of nuance to using this, and it will take a while to get used to.

Once you're used to it I think it could work, but learning how to use the controls is a big deal. For the most part I can control the stereo or temperature in any car that I may happen drive, and so can most people without really much of a learning curve. Setting the clock or mucking with a GPS might be a bit tougher, but the basics are all there. The most likely time that you'll be learning how to do this kind of stuff are while driving, so it should be as obvious as possible, with as little need to focus on it while figuring out the controls.

Cool. Useless. As Don Norman might say about it years from now "it probably won an award".

I have a simpler and -- I think -- demonstrably better solution:

1) A minimalist HUD overlaid on the windscreen. (As simple as possible -- an icon, text label, and a simple readout, e.g. "Nav Icon: Driver Side Climate 72°F" or "Sepulchre Ave <-- 500ft" or "FM Radio: WAMU 88.5MHz [xxx ]")

Benefit: Driver's eyes stay on the road. What you're doing is utterly unambiguous.

2) Steering wheel buttons that select the mode (channel select) and value (volume buttons). For bonus points, allow two value selectors (so you can adjust station and volume in one mode

Benefit: Driver's hands stay on the road. Steering wheel has touch-affordances to help you find things on it. The buttons already exist on many (most?) cars.

This seems like a really clever implementation of an system that's probably a passing fad.

Touch screens are great for portable computers both because they need to be small and because they need to enable a potentially limitless number of operations.

In a car the number of things that need adjustment while driving is relatively small and doesn't really change. We've had climate and audio controls since the 1950s. We've since added speakerphones and navigation, but the former only requires a couple of buttons, and the latter shouldn't really require any adjustment while driving.

There's also no shortage of space to put the controls.

My guess is that touch screens will start to dominate in cost-sensitive segments of the industry, while physical controls will migrate to luxury status in a few years.

I'm actually a big fan of the Hyundai Veloster's touch screen. It provides all the functionality you'd expect, without the convoluted menu systems of other cars.

http://gaspuol.files.wordpress.com/2012/11/2013-hyundai-velo...

http://www.extremetech.com/wp-content/uploads/2011/09/MY12_F...

Granted, this UI looks great, although has a few interaction oversights: what if I lost a couple fingers in an industrial accident and only have 3?

BMW's iDrive was/is considered a mess because of the single knob-do-everything approach, but it had potential for what could had been. This seems like it took what bmw had, implemented it as a virtual knob and removed the main problem (idrive's submenu system). I like this, great way to think about the problem and present a potential, easy to understand solution. This approach also allows the touchscreen to be put in a more natural position like closer to the arm rest. I'd just make it a taste more visual. I bet the haptic feedback that some android keyboards offer when keys are pressed would be a great addition.

Lovely, but one fairly critical problem - for the 88% of people who are right handed, driving in the 35% of the driving world that drives on the left, they'll need to operate the centre console with their non-dominant hand.

Some things just need buttons.

Count me in the "give me my knobs" crowd: for frequently-accessed features, I absolutely want physical controls. Including temperature and fan, audio, AC, etc. I'm not even much of a fan of the shift toward "programmable" temperature controls: when I'm changing the temperature, it's because I'm too hot or too cold -- what I want is to move the indicator toward where I want to be and not fuss with it. A given cabin temperature: 68F, say, might be comfortable on an overcast day, too warm in direct sunlight, or cold when it's below freezing outside and drafts are intruding.

That said, a few nice features here:

• Moving the control to the user. Ages ago I noticed a trends in desktop UI. When a window would open a dialog, typically it appears more-or-less at some random location on screen. Microsoft came up with a solution for this: they'd "warp" the mouse pointer to the dialog location. Which meant that wherever I thought my pointer was, it wasn't. A crufty old Unix graphics package, xv, had a much more elegant solution: It would open the dialog, with the default option button positioned under the pointer. It's subtle and it took me a while to realize it, but it floored me when I finally caught on. Sadly, I don't know of any other application (much less desktop or GUI) which practices this principle. Matt's example here does, it's the first I've seen of this in nearly two decades.

• The multi-finger touch thing ... I might actually get used to that. Maybe not for primary controls as indicated, but for other advanced features (say, a mapping / navigation system). No, it's not going to work for everyone, but a fallback mode (or modes) could work: hot corners, or similar features.

• Keeping the interface pared down to just what's being acted on is useful. I've been going through the process of stripping down web interfaces, mostly for my own benefit, and one thing I've come to realize is how an ugly, messy, disorganized UI often is hugely improved by junking most of it. For most websites I visit, what I'm interested in is the primary content, so getting rid of the ancillary bits doesn't merely lose little, it adds to the experience. For UIs in complex spaces (and a car's internal systems are a modestly complex example), you'd want to be able to dive into a richer interface from time to time, but getting to the bare essentials is definitely useful.

So while I'm not sold on the specific implementation (mostly the use case), the concepts here are good.

My favourite touch screen UI on current cars is the one on the VW Golf 7. It has this proximity sensor that highlights and increases the size of buttons when your hand approaches the screen.

Overall, the interface is very clean and avoids a lot of visual distractions present in many of these UIs.

It lacks the responsiveness/speed we see on modern tablets, tough, this seems to be the rule on the market, even Tesla's screen is quite slow.

Video: http://www.youtube.com/watch?v=GyBUcCno_i8

One crazy person with a great idea really has a chance to make it in today's world. This entire video and app could be made by 1 person to demonstrate the power of their ideas.

I love that we live in a generation where we have the tools to display our ideas - without the concerns for cost one would have endured trying to build a prototype.

Someone might have an idea for a car dashboard – which, until recently, would have had to include many other people - suppliers, plastic molds, etc. Today we have 3D printers that can facilitate this process.

GM did a touchscreen back in the late 80's: they embedded a small green screen CRT in the dashboard of early model Buick Reattas. The general consensus at the time was that it was distracting and hard to use. Two model years later, the Reatta switched to physical buttons.

It amazes me that so much of the modern car industry seems dead set on repeating a 25 year old GM mistake. My wife and I have a 2013 Mazda CX-9 with a touch screen, and it is exactly as difficult to use as the 1989 GM model.

Definitely in the right direction. I've never thought long about solving for car interfaces, but I'd probably keep a more visual interface around and build it with these shortcuts in mind -- a basic point and touch interface for new users, and more gesture-based control for people that learn the system.

I'd also probably not do more than three-finger combinations. One, two, and three finger controls are pretty easy. Four and five require some hand contortions on car-sized screens.

Sliding is nice for analog input, for 'discrete' input it needs even more radical schemes, like fitz infinite-borders, NESW or 'hot corners'. And with slight visual cues like thresold lines too.

ps: funny to see videos about sensitive volume control when this was on front page https://bugzilla.gnome.org/show_bug.cgi?id=650371 (gnome hardcoded volume steps)

I'm sure it's been talked to death but touchscreens lack the most important feature of any car interface. The ability to use it without looking at it.

Maybe he isn't solving this problem but I'm glad it's being thought about. 10 years ago I think swipe to unlock could have the same crticisms. People can and do adapt. There eventually will come a set of standard gestures that make sense for most situations. These will become standardized and everyone will think that this is the way it should always be done- until someone else creates the next big thing...

I love those UI. They bring that sensation of abstraction and less Remote Control Syndrome. The only problem I saw in my tests with real people is that those kind of UI have a higher learning curve.

People need to learn well all the actions, even if they're 5 and learn the steps well to achieve actions. It's different when they have 100 buttons on the screen, the only thing they need is read lineary and touch.

It's really nice concept!

I really like the “one big button” idea!

Using different numbers of fingers to control different things is interesting- how would people with missing fingers use this?

Some Lexus vehicles I've been in featured a center console joystick that vibrated when the cursor moved over an element on the screen. I found it to be the most intuitive and the best at keeping my eyes on the road while still being able to navigate the various menus. Not perfect, but certainly better than a touch screen with no feedback requiring you to take your vision off the road.

Neat and I feel like this could be easily learned by a driver. Minor feedback: the audio level selection seems to move on linear scale, but volume is a logarithmic scale. The perceived difference between 16db and 8db is the same as the perceived difference between 4db and 2db. If the audio is on a linear scale, then the sensitivity of the screen will be much more on one side than the other.

I guess all of America will just memorize these finger combos?

Teach them in schools? Remember kids when confronted with a blank screen 3 fingers is audio source!

Looking over this again, two things strike me:

1) Why was a single touch interaction not demonstrated? Each control requires 2+ fingers. Maybe that can bring up a help menu or other information?

2) There is no demonstration of boolean switches and button controls. Things such as turning on a defroster, or enabling pairing with a bluetooth device. I'd be curious to see how those would play into the ux flow.

Touch screen in cars? I'd prefer to keep my eyes on the road, thank you. The screens have a bit more things to control than I can remember. I guess my muscles have terrible memory. I can't imagine how so many features can be controlled using memory-friendly gestures and whatnot. The article and video only show that it's not possible.

- there's a learning curve - that problem was solved a long time ago by putting the most important buttons on the wheel. - the car screens are not as big as the iPad, at the moment you put all your 5 fingers, you might occupy the whole area, and you won't have anywhere to scroll

given all that, that video proposes a new approach, not a solution.

Some people have already mentioned that you may want more than 8 controls. Perhaps a solution for this would be to have a max of 8 controls but the user could define which ones are mapped. Perhaps this interface can be a layer on top of a more traditional interface. That way the driver could customize the controls to fit his needs.

I am betting on voice. I don't think you should be looking at a screen while driving.

Http://plexinlp.com is building an eyes off interface for cars (and other things) to me the ui should work while looking at it or via voice. Just like conversation s with humans. They work over the phone, but they can be reinforced with visuals.

What's wrong with people pushing touch interfaces in cars? I'm still using old Garmin StreetPilot 2620 for navigation because it has hardware buttons which I can press without looking at them. Almost all later navigation hardware is touchscreen only, which requires to look at them not at the road.

Seems interesting as a study, but what about people without 5 fingers?

Do they get 20% less things they can control?

I ask in all seriousness because I know several people who have lost one or more fingers.

It seems to me that something like Google or Apple voice commands are what is really needed, assuming on wheel controls don't suffice.

There is going to be a lot of hate in here about how impractical this is, but I just wanted to say nice job on the clean design and presentation. I watched through all of it.

I wish more car touch-screens were capacitive.

Maybe having a physical button/dial with these features (like the Nest thermostat) would give people more connection to the 'past interfaces'?

You still down know if the display registered the correct amount of fingers. And you have to look to check.

I originally posted this as a "Show HN". Does HN update titles automatically sometimes?

Touch screens in cars are a great idea... in places with smooth, pothole-free roads :-)

This builds on the basic idea behind pie menus, which is to exploit Fitt's Law, by lowering the target distance by centering the menu around the point of contact, and increasing the target area by basing the selection on the direction of movement instead of hitting a small absolutely positioned target. https://en.wikipedia.org/wiki/Fitts_law

An Empirical Comparison of Pie vs. Linear Menus, Jack Callahan, Don Hopkins, Mark Weiser and Ben Shneiderman. Computer Science Department University of Maryland College Park, Maryland 20742. Presented at ACM CHI'88 Conference, Washington DC, 1988. http://www.donhopkins.com/drupal/node/100

I've applied pie menus to various applications including tools like window managers and text editors, and games like SimCity and The Sims, which uses pie menus for controlling the lives of the simulated people. https://en.wikipedia.org/wiki/Pie_menu

Since the direction of motion, which selects the pie menu item, is independent from the distance of motion, you can use the distance as a parameter. Here's a demo and an article showing various kinds of pie menus I've developed:

Pie menus from "All The Widgets" CHI'90 Special Isssue #57 ACM SIGGRAPH Video Review. Including Doug Engelbart's NLS demo and the credits. Tape produced by and narrated by Brad Meyers. Research performed under the direction of Mark Weiser and Ben Shneiderman. Pie menus developed and demonstrated by Don Hopkins. http://www.donhopkins.com/home/movies/AllTheWidgets.mov

The Design and Implementation of Pie Menus -- Dr. Dobb's Journal, Dec. 1991. There're Fast, Easy, and Self-Revealing. Copyright (C) 1991 by Don Hopkins. Originally published in Dr. Dobb's Journal, Dec. 1991, lead cover story, user interface issue. http://www.donhopkins.com/drupal/node/98

Here's a more modern implementation in Unity3D: http://www.youtube.com/watch?v=sMN1LQ7qx9g

The multi finger interface is obviously a very nice touch that this interface brings to the tablet (if you'll pardon the pun ;). I like to call touch screen pie menus "finger pies", in a nod to the Liverpool slang term from the Beatles' "Penny Lane": "a four of fish and finger pie". ;) http://www.songfacts.com/detail.php?id=115

"ConnectedTV" is a finger controlled Palm app that David Levitt and I developed more than a decade ago for the Palm (without multitouch of course): a handheld personalized TV guide integrated with a universal remote control, which you could reliably operate with your fingers instead of the stylus that most Palm apps required at the time. Here is a review than mentions the single handed finger touching and stroking interface:

ConnectedTV: TV Guide + Remote Control, Geoff Walker, 2/2/02. http://www.pencomputing.com/palm/Pen44/connectedTV.html

"On the Palm, the application user interface is simple. The application checks the clock on the Palm and displays the names of all the TV programs for the current 15-minute window. The program buttons are large enough to easily hit with your finger. For easy identification of programs that are about to start, one corner of the button is clipped. To tune the TV to a selected program, you just tap the button. To display more information about a program (i.e., the blurb from the guide), you do a downstroke on the button. The downstroke is a result of the program menu buttons being implemented via "pie menus" (see www.piemenu.com), which makes it very easy to use the application with one hand."

"For instant access to frequently used functions, the application remaps the Palm’s hardware buttons to Power, Mute, Volume Up/Down, Next and Back. At initial release, the program will include support for TiVo and some brands & models of VCR, as well as the ability to program it for other brands. Management of "Favorites" (e.g., flagging favorite programs with their first run/rerun status) is planned but may not make the first release."

Most Palm apps required a stylus to operate, because their interfaces required a lot of precision and visual feedback. But it was quite easy to accidentally lose your stylus in the couch cushions in the dark living room, so pie menus made it possible to control the Palm in a dark room with your fingers, which was unusual at the time (long before the iPhone).

ConnectedTV used big multi-purpose buttons that you could reliably operate with your finger, supporting up to five complementary functions for taps and strokes up, down left and right, with immediate audio and visual feedback during tracking.

The directional strokes worked very nicely with complementary sets of commands that remote controls tend to use like changing volume up and down, changing channels next and previous, liking and ignoring programs, paging up and down, moving forwards and backwards in time, etc.

David Levitt wrote up a summary of how it worked:

iPhone Lovefest - How Don and I Invented Stroking vs Poking https://www.facebook.com/note.php?note_id=106220169912

On a recent Sunday, journalist Steven Levy hosted an iPhone Lovefest at Sylvia Paull's Berkeley Salon, and more than a hundred of us showed up. I realized some of us had quietly helped invent something historic, and I spoke briefly about it that day, but first I did a little demo.

More than a dozen developers brought iPhone applications to show off; a camera and projector showed the crowd each little screen. I went up last, commenting that with its sensor, the iPhone had the same motion sensitivity as the Wii's legendary game controller, and tapped mine on the table. As it tumbled from my hand there was a nasty cracking noise and the audience saw this: Gasps and shrieks were soon followed by laughter, a woman shouting "I have version 1.0 of that!", and relieved applause. I explained that the iPhone application draft I was showing was called iBustedIt, the cracking sound and image were triggered by motion, and Levity Novelty aimed to make it available by April Fool's Day. Only it turns out Apple considers this kind of trick an impersonation of its fine software and doesn't allow it in the App Store - so my apologies to those of you who planned to trick and shock your friends with iBustedIt. (Eventually I learned Ms "I have 1.0!" is Ann Greenberg, whose iPhone glass is actually cracked that way and still works fine.)

As the laughs died down, I noted how easily we can miss major technology transitions, not realizing what's actually new. When I developed software for the Palm in 2002, every handheld application required a stylus: to operate the tiny on-screen buttons you had to use both hands and poke them with a little stick. Remember?

Back when my old company ConnectedMedia created ConnectedTV - a Palm app showing personal TV listings that let you change the TV channel by touching the screen - I considered the poking interface unacceptable. Requiring a stylus would mean a nerds-only experience no consumer could love. Unless you could operate it with a thumb - and with one hand - it could never compete with a remote control.

So the on-screen buttons had to be much larger, with each screen elegant and simple. Only, most screen space was needed for TV listings and show descriptions. My ingenious friend and partner Don Hopkins realized that we could repurpose the Palm OS screen 'stroke' detection capability - previously used only as a way of entering characters - into a kind of short cut: if you stroked the button, its label would change with a click, and then if you lifted your finger it would perform the function shown. Since you could easily stroke up|down|left|right, we could stash up to 4 helpful shortcuts under each finger-sized screen button.

To prevent accidental channel changes, we also required you to stroke down on the TV show name or description to watch it. It worked perfectly. Of course, this is a predecessor of the swipe today's iPhone needs so it can't be unlocked accidentally as you carry it.

Users adored ConnectedTV, calling it "addictive". Handheld Computing called it "one of the most impressive Palm OS applications we've ever seen." It was more responsive, more personalized (stroke to set a Favorite show), and less obtrusive for browsing than the TV's on-screen guide.

Sony paid us to make custom 'skins' for their sweet CLIÉ line of Palms for cable TV trade shows, and then offered to bundle ConnectedTV with the CLIÉs, pre-installed. However, soon Sony Japan dropped Palm OS and discontinued the whole CLIÉ line. Still, our stylus-free interface had provided a peek at the strokable handheld future.

Stroke replaces Poke Just a few years later, no popular handheld device uses a stylus. The iPhone has expanded the idea further to support animated scrolling and multi-touch pinching.

Is there really such a big difference between poking with a stick - occupying both hands - versus being able to stroke, rub, caress with a finger, and even pinch? Um, ask your mate. Such quiet GUI software innovations help us fully enjoy, literally embrace and yes, love our technology.

why the need to touch a screen? In a car, wouldn't it be better to have gestural controls that don't require looking at a screen or touching it, or voice control?

What if you don't have five fingers to work with?

Trying to be cute and kill the driver... Nice work!

Now I have to change my interview question. :(

Real buttons are cool. /thread

Or, you know... physical buttons.

Touch screen in cars... Nope

This is like learning vim!

There's so much to say about this video, but I'll summarize it by quoting one of my favourite articles: "Solving the wrong problem". [1]

(By the way, itt's funny how nowadays, hardware UI is immediately associated with "touch controls".)

[1] http://prog21.dadgum.com/130.html

I prefer to talk to my car computer, it's based on Windows CE [1]. It works like Siri and predates it by several years.

http://en.wikipedia.org/wiki/Windows_Embedded_Automotive

I think touchscreens are a thing of the past and we should really look forward to a new, more intuitive technology of the future. I've been thinking a on and off about this and I think I have the solution, though it may take a company as visionary as Apple to implement.

What I've come up with is a system called finger-friends. Where the touchscreen usually goes, you'd have a grid of finger-friends and some information-displaying screens that do not react to touch (this is important). The actual finger-friends are physical objects that can be manipulated by a person's fingers and relay control intent to the software. So far I've come up with four (well three, you'll see in a second) finger-friend designs to replace our usual buttons and sliders.

First is the pushy - it's like a button, but instead of reacting on touch, you need to push it - it's basically a nub with a spring inside that closes a circuit when it's pushed to the bottom. The good thing about this is that you know when you've pushed it to the bottom and don't need to look at the system to know it's accepted your touch.

Second is a variation on the pushy - the sticky pushy. Like a toggle button, it has two states and works by having a latch mechanism - you push it to the bottom and it stays there, and when you push it again, it pops out. I call the two states of the sticky pushy pushed and popped, and propose that pushed state should be used for "on" and popped for "off".

Third, the tweaky. It's a protruding cylinder with ribs along the ridge for easy grasping. You turn it left or right to increase or decrease a value, like a normal dial. It has a leftmost state past which it can't be turned any more and a rightmost state. It also has a line painted along the edge to show which way it's oriented and you can paint the range of values around the base so the user can see which value in range is selected.

Fourth is the snappy. It's a small stick pointing outwards from the dashboard moving inside a ridged groove which lets you set it at one of several discrete inclinations. That is, when you push on it from the side with your thumb it snaps off its current ridge and snaps into the next one. This allows you to select one of several mutually-exclusive options.

There might be other finger-friends you can come up with, like something to replace the normal radio button controls - the snappy can function as that, but I'm worried its visual language will confuse users and they won't see it as an actual radio button group.

Another problem is how the information you can put around these controls is static. You can't make them magically show as actual physical objects in the middle of a normal touchscreen, so maybe the screen should be broken up in several pieces and each part adorned with finger-friends around the rim. The screen would show information on what each finger-friend does and allow you to switch your mode. So, the most used features in the car, like the headlights, the turn signal, the heating and the wipers would have their own dedicated finger-friends along the top of the dashboard with their one function written on them, while less-used ones will be bound to contextual controls explained by a screen.