We're obsessed with creating interfaces, and putting them everywhere. Let's call it the "put a bird on it" effect: slap a UI on something (a refrigerator, a car dashboard, a hotel lobby) and presto! Instantly better. But @goldenkrishna clearly shows how this has taken a turn toward the absurd. By chasing the next screen-based solution, UX is heading down the wrong path. Meanwhile, Pay with Square and others are quietly (or not so quietly) showing how No UI solutions can deliver "natural" experiences without tapping or clicking.

Judging by the response from the twitter-sphere, this clearly hit a nerve. Why? What about it feels so right? And where do we go from here?

I think this is an example of leading with technology and of creating a solution in search of a problem. It seems like a fad-ish trend to add a screen to everything.

The automobile industry is adding touchscreens to dashboards, and customers and experts are complaining about diverting attention away from the road and creating yet another UI to be learned. We've had a volume control knob on radios for decades - do we really suddenly need to (a) tap to activate a touch screen, (b) swipe to find a radio app icon, (c) tap to open it, (d) visually scan for "Volume", (e) tap to open a volume control, (f) swipe / drag while watching the screen to change the volume, (g) tap to close the radio app, and (h) tap to return to the home screen?

"Motor memory" is real. We learn to DO things with motor skills (and tactile or haptic feedback) that can be more efficiently and effectively performed with motion than with thinking. It seems counter-intuitive, but once a motor skill is learned, it can be executed with drastically reduced attention (thereby freeing up cognition for more important things.)

There is a reason the cockpits of airplanes are (a) standardized, and (b) still full of switches, knobs, and buttons. Those mechanical bits keep the pilot's attention where it should be (not scanning for app icons) and to reduce performance time and errors.

Screens are not a solution to every problem. They are a potential solution to some problems.

A major promise of technology is to do more with less effort. Sometimes the "less effort" part takes a backseat or is utterly ignored. Dave's automobile touchscreen flow description is a perfect example.

This is where No UI comes in. It's a spotlight on the importance of reducing the effort we devote to mundane tasks so we can redirect it to important endeavors.

@DaveHogue "a solution in search of a problem". Nice. OMG my fridge doesn't have an app! It does, actually: to keep things cold. And it already does that.

The physical switch angle is interesting, too, though wouldn't it be great to have no screens AND no switches? Obviously, mission-critical environments like airplanes need to be more conservative in this regard, but when you're paying for a coffee, why put more UI in the way?

I'm not sure that "No UI" is actually what we are talking about, in a strict sense. Isn't it really about best matching the modality of the UI to the task or problem at hand? A knob, switch, or button is a mechanical UI. A mouse is a mechanical UI that translates motion into digital data. A touchscreen is more direct manipulation of information in a digital form. Spatial gestures require detectors. Voice recognition requires microphones.

UIs are really just the sense organs of machines. They are essential to detect and translate real, physical forces into digital data just as our sense organs detect and translate the physical work around us into neural / electrochemical signals.

Efficiency seems to come when our "sense modalities" match.

Yes, you're right, strictly speaking: sensors (being the sense organs of machines) are required to get machines to understand the world, including you. That makes them an interface. But in practice, you can imagine designing experiences around people that require no (conscious) interaction with those interfaces at all. Walk up to your car, the door unlocks, you open the door. Order a coffee, you get the coffee, you drink the coffee.

The "No UI" idea is a short-hand for the notion that there are a whole slew of behaviors for which a direct UI (on a screen, a physical object, whatever) is entirely unnecessary, replaced by a bunch of ambient always-on sensors that just do the right thing without being asked or told.

Ah, I love the idea of the "Ambient UI" and "ubiquitous atmospheric sensors" (though there seems to be something creepily Orwellian about that...) There /is/ a UI, just not one that requires our direct interaction (or even awareness), but we could still manipulate the outcomes if we chose to.

For example, automatic doors and proximity sensors: they notice our approach and open the door ahead of us. People with arms full of groceries easily walk right through, barely giving a thought to the convenience. However, we can make ourselves aware of them, and we can attempt to thwart to divert them. Have you ever watched small children try to sneak up on an automatic door? They clearly are interacting with a passive, ambient UI. :)

Oh man, fooling the motion sensors: one of my favorite ways to pass the time at the grocery checkout line as a kid.

You've hit on a key objection to this overall thread, though: what of security and privacy? Maybe someone will dream up an https equivalent for location or proximity.

Creepy: when it does something strange.
Magic: when it helps us.

I think we could easily go a few steps past the automatically sliding door.

Beyond simple sensors...temperature, location, microphones...big data and pattern-seeking AI could help people find the solutions to their daily problems instead of forcing screens upon people who don't want or need them.

"Any sufficiently advanced technology is indistinguishable from magic." - Arthur C. Clarke

Expectation: a desirable action or outcome to which we pay little or not attention.
Magic: a desirable action or outcome that we notice but do not understand.
Creepy: an undesirable action or outcome that we notice but do not understand.

There's a matrix somewhere in there with axes for: attention, desirability, and understanding. What would be in the other cells?

How do we know when to make the system noticeable and help people find solutions versus keep the system below the threshold of attention and simply perform the solutions for people?

When do we need to involve people? When should be intervene on behalf of people?

How many good decisions does a machine need to make before people will trust it (implicitly) and no longer prefer involvement?

I love "No UI" as a buzzphrase, but I'd argue that a GUI is not always a problem. There are more examples of "too much GUI" or "wrong GUI," and I think the fruitful conversation is really around "very little GUI" and "appropriate GUI."

I would even argue that you could reasonably put a GUI on a fridge. Fridges will be smart, and perhaps next-gen fridge designers want to ensure that you never worry about forgetting to buy milk when it runs low. (Talk about first world problems). There's a "No UI" solution to this problem: Tie the fridge into a grocery delivery service, etc. But a simpler (and arguably more human) solution would simply allow the fridge to helpfully inform your shopping list, right? And you could imagine a mobile GUI, and even a fridge-based GUI for this, right?

twitter.com

@douglemoine: The auto-mobile shopping list seems way more compelling than an LCD on the fridge door. I want something to respond to the fact that my milk is sour, but why would it be a screen on the fridge?

I get a box of veggies from a CSA every two weeks. I don't know what's going to be in it, but there's always a lot of it, and it's seasonal and tasty. And it comes with recipes, and a dozen eggs. Boom. Done. I don't even have to make a shopping list. Never in the process (except on sign-up) do I have to use an app to confirm my purchase of a dozen eggs. Talk about reducing cognitive load.

Anyway, you're right that the REAL question is one of "appropriate UI". "No UI" is simply a vastly under-utilized category of solutions.

No matter how cool your interface is, less of it would be better.

Aza Raskin has an interesting blog post about efficiency, and this can be tied into the idea that less is more:

azarask.in

Another (more obvious) connection is the paradox of choice: Fewer (good) choices is better than more.

The Victorians embraced "more is better" (just look at their architecture and interior design!), and Mae West said, "Too much of a good thing is wonderful!", but we UX folks should have tools and techniques to help us identify when we've gone too far. How do you know when you've crossed that line?

Wow, Dave, that aza raskin article is intense. Great new vocabulary to talk about this... Get rid of 0% efficiency UI, of which there is more than we commonly acknowledge.

"If there is only one next logical step, have the computer do it" – I think that's a Cooper quote from About Face.

A few people have related "the best interface is no interface" essay to Scott Jenson's "Just-in-Time interactions"...thoughts?

I think that it is such an interesting time and movement to be a part of. I admire a lot of the contextual awareness things happening right now, and I certainly think Pay with Square and others have shown how useful it can be without an interface.

I do think that a constant reminder of less is more is always useful when building an interface for an application.

I love this thread, and hope it continues—thanks for including me!

@michaelbrooksjr - you should get on the mailing list for the Cooper Parlor, especially if you're in the SF area. @goldenkrishna will be keeping the conversation going in person. Do that here: form.jotform.us

And @goldenkrishna: the JIT interactions piece shares some of your frustrations with the _weight_ of apps - downloading them, finding them on your phone, opening them - but doesn't acknowledge the fact that _using_ them also carries a burden. It still forces an interaction where none is necessarily indicated.

I am interested in people's perspective on systems that learn from our interactions with them (the Nest thermostat being one example.) What if every time there is an interaction with a system, that system in combination with some context (time, location, temperature...) learn something about my usage pattern and adapts accordingly.

On the notion of ambient sensors, Teehan + Lax created a simple service that would passively monitor how much milk you had left and send you a quick message when it was running low (and it also shows you the nearest store). The demo video's worth watching... teehanlax.com

@burcbiz I think this is the direction we are headed in, and it could provide a lifetime of opportunities to master it. I'm looking forward to being a part of the movement.

I have been thinking about this for the last two years, but I haven't been able to wrap the ideas neatly, ending up with more questions than answers.
Will it be "creepy" to the user and/or more confusing?
Are there general solutions, or each does each system need its own solution?
Does the interface disappear or become optimized?
Is the interface designed for first time users only? (As the system learns we will interact less with it.)
Assuming the user still needs to interact with the system, is the goal to "asymptotically" become %100 efficient (using the razkin vocabulary.)
Once in learned mode, will it make it harder to execute new interactions?
…

@Nate I see what we are getting into, we are "Interfacing" in a wrong way, but I don't think people do this because wants to mess with customers, example: I was hired to build an app for Mexican Government, the purpose was to allow citizens check the mexican grocery list prices in their markets so they can go the nearest and cheapest.
The manager want this workflow: user opens app -> select Mexican State -> select market -> search for a product -> select product -> check price

And I said nope, why should we ask for things we already know, let the smartphone check where we are, lets make the common grocery list prices and let the app check where is the nearest and cheapest market. They didn't realised about the smartphone's capabilities.

My point is, as we understands technology's capabilities and limitations we will be able to design better products, perhaps these products have a minimum interface like Nest, perhaps not, but certainly it will have the necessary to interact with in a good non intrusive way. I don't think that going wild and try to make every single product "interfaceless" (note the quotes) would help.

By now, how to achieve this level of understanding?... well I think only experience and starting to make design patterns for our day work. The above example I called "Service by Location" where according to a location you can offer tickets, places, deals, etc, etc. maybe it's time to start compelling what we have learned and try to achieve the TBIINI goal.

I am not fully-settled on this, but sorta think of the interfaces that support IxD as not quantum: there or not. There's a continuum. Aside from the degree to which the on-screen items get in your way, I mean that there are many interfaces. Here's an example list of things I have output user info to (excluding desktop and mobile web and apps), just as an example:
• SMS, MMS
• Push messaging
• Email
• IVR
• People - Scripts and call center stuff to help them communicate better
• Stickers - On the box or product to explain or guide
• Containers - The box you store stuff in can be designed to support the use of the contents
• Annunciators - Lots of lights are not useful. Is red bad or just "on"?

It's also, very often, several of these.

(We may be getting a lot of new participants in this conversation - please feel free to use the "branch" functionality and create a new side thread from someone's existing post if you want to have a more targeted conversation or tangent.)

I'm really excited about anthropomorphic interfaces as a vehicle to get people to elicit their intentions sans traditional UIs. While Siri, et al, haven't nailed it yet, I believe there's a huge potential for people to explicitly ask for what they're looking for -- Not just for basic information retreival as Apple and Google would have you believe, but a triage mechanism for more complex requests that people have.

Full disclosure: We're trying to explore what this means by creating a simple chat-based interface that's a personal assistant who helps you schedule meetings with other people.

Check it out at Lizi.ai -- We'd love your feedback.

An expression of the no-UX approach? I had needed this, and I like the approach. "iOS 6: Messages Now Automatically Selects Last-Used International Keyboard For Each Contact"

@Golden Krishna - I stepped away from this thread for a little while, so sorry for the delay.

In education and professional development there is the comparison of "just in time" vs "just in case" training. It is far more effective to teach someone what they need at the moment they need it (permits instant use/application and enhances memory) than to train someone on something they MAY need at some undetermined point in the future (delays use/application and does not support memory.)

Many apps (mobile, desktop, anywhere) offer tons of functionality just waiting for US to recognize the moment we need it. Maybe "No UI" also means "get out of the way until needed, then present only what is needed and hide the rest": Context awareness.

@Burc - I really like your thoughts on systems that learn, designing for developing expertise (and how the UI may change with our skill level), and ambient input (sensors, UIs) and output (feedback.)

I've been watching Google Now (in Jelly Bean, Android 4.1), because I think it may be trying to learn from my interactions. It is context aware (what time is it? what day is it? am I at home? at work? traveling? if I search for a specific flight, it adds a flight tracking card...) and tries to ask me a minimal number of set up questions (e.g., do I drive or take public transportation?)

Google Now is an first step toward person-matched machine learning; I think my phone is learning about me, and it has not yet entered the uncanny valley.

One of my co-workers ran across an interesting No UI example this morning. It doesn't express principles of learning, but it does embrace the natural process of approaching a door you want to open. There's also some UI to get things started, but it's fascinating nonetheless.

theverge.com

Deviating a little. Fortunately in the web world the trend is towards simplified and sharper interfaces. Long gone are the days of (usually poor) flash sites, or those built with excessive imagery. There are a lot of relics out there though.

Look at the fairly spartan colours and layout right here on Branch. A simple sidebar and the content. And a few links. That's about it.

Or the new microsoft homepage. Sharp and uncluttered. Even apple.com, though it needs a bit of an update to make it responsive to mobile devices. Ironic they ship iphones but their own site looks poor on the device.

For a poor example look no further than facebook. What a mess of an interface. The mobile app versions are somewhat better.

@Al, but this simplicity comes at the cost of features no? Take branch as an example, no search, no browsing, just a few categories on the home page you can't control.

You need to analyse your site or business and decide which features your users need of course

"I think we are on the cusp of an important transition within IR #3. So far, we have used these innovations to make ever faster, smaller, and more useful computers, including mobile phones. We have created, as Gordon notes, a whole lot of dot-coms and online services. But we’re already starting to see engineers and companies dabble with new kinds of products. Rather than merely accepting, transforming, relaying, and displaying information, some new computer-based products have more of a physical—really, a kinetic—effect on the world."
- Eli Dourado

elidourado.com