Hacker News Reader

He was wrong then, and he's wrong now. We know he is wrong because computers operate fine with 101 key keyboards just like they did 30 years ago. The million of applications produced since haven't required a new button on the keyboard every time.

Thankfully some sanity has reigned in car interiors and the manufacturers have realised the benefits to tactile feedback. He might be right when screens can reconfigure themselves with raised sections. Or mind control I guess.

It's all well and good saying he was wrong but the principles he outlines have been largely vindicated by the market.

Many things worked fine 30 years ago and yet have been immeasurably improved since - it's a tempting argument to not fix that which isn't broken, but ultimately a weak one. Also, there is little point in comparing computers, phones and car interiors. Incorporating physical buttons on phones involves a significant trade-off that is not an issue with computers. Car interiors are a different matter entirely. A more appropriate comparison might be conventional TV remotes vs the Apple TV remote. I've found the latter to be much more user friendly and intuitive.

> it's a tempting argument to not fix that which isn't broken, but ultimately a weak one.

I'm going to have to call shenanigans on this one. Sure, it's an alright attitude to have towards non-essential parts of your life/consumer experience, but imagine if mission critical stuff (like, flight consoles) changed on a regular basis - that would be completely insane.

Fortunately they don't, and fortunately neither do smartphones. They've been roughly the same for the past 12 years, right around the time people started to actually use them en masse. People are already adept at understanding how to use the on-screen keyboard when it changes, such as when it only requires numerical input, or when filling out an email address it helpfully puts the at symbol on the main screen rather than behind a shortcut, or more recent changes like buttons to suggest filling in a saved password for an app or website. These are barely hardships on the scale of unfamiliar flight controls for experienced pilots.

Car interiors are a wildly different use case. The safety concerns of operating a flat surface while driving and having to keep one's eyes on the road probably outweigh the benefits of a flexible UI design. In practice, however, the best approach seems to be a hybrid one: Keep basic, important functions physical, tactile, easily and blindly accessable while serving the details on a blank canvas.

For phones and tablets I prefer the giant screen approach.

I feel when people can make such blanket statements about what's right or wrong, they are only thinking about their own use case. Perfect example of what Jobs was talking about: the 101 keyboard works fine for the applications designed around it, but there are plenty of things for which something more optimised is a better fit. The argument about millions of applications not requiring specific buttons is a bit fatuous; there are plenty of examples of times when a physical keyboard could be insufficient for efficiently solving a problem on a smartphone without resorting to inventing a strawman set of millions of applications irrelevant to the specific domain of user input experience. I'll give you one example:

A lot of HN users seem only to write in English. I write in three languages, one of which is very different from the other two. Using one keyboard is possible on a PC, but would make for a lousy experience on a smartphone where the ideal layout for high speed typing in that latter language has nothing to do with the typical keyboard layout. To transplant this language's layout on to a shrunken down imitation of a full keyboard would be a productivity loss. This is a scenario where a physical keyboard on the phone would be a hinderance, not a help.

And that's not even considering many languages that are similar, use the same alphabet, or whatnot. They might be able to get away with the same physical layout, but now the manufacturers have to decide whether to print the legends on or near the keys with the standard US layout, UK layout, French AZERTY layout, and other layouts specific to each region, sometimes at great expense for relatively small markets.

That is if they can fit all the legends in. Some languages will still need a visible set of Roman characters, so not only will the local language's characters be printed on or near the letters, but so will the standard QWERTY layout, taking up valuable physical space, possibly necessitating a smaller phone.

I'm not mentioning what specific language I used that's quicker to type on a keyboard layout other than the standard PC one because it doesn't matter: it is within our capabilities to not have to manufacture a hundred different configurations of physical keyboard for smartphones and tablets for years, long before the arrival of the iPhone and iPad. Nobody needs to know what language it is because it doesn't matter: those who don't even vaguely consider the needs of international audiences as being first class, on the same level as the needs of US English users, don't have even the beginning of an understanding of designing products for international audiences.

The whole /world/ uses smartphones with touch screens. Not just business users, not just executives on meandering lunch breaks with their Blackberries; every day actual people who don't want to be bothered with which SKU to get in order to fully support their language, who don't touch type on regular keyboards let alone tiny ones, children, grandparents, and everything in between.

That's cool, but I think that most people here agree that mice are a pretty terrible method of interfacing with a computer. Every time you use a mouse, you have to re-orient your self to the state of the mouse (its current location and its use within a given application). That is why we spend so much time optimizing applications for use with a fixed controller (a keyboard).

A touch screen is a bit better than a mouse for lots of stuff and it preserves the discoverability of a mouse-based UI. Unfortunately it retains the drawbacks of a mouse based ui. Worse is that there is no fixed controller on a smartphone that might allow faster interaction w/ apps. This is where I think we are getting in trouble. In a traditional computer setup, you interact with the mouse when you are learning new software, and eventually transition to using a keyboard more and more. With touch based stuff, it seems like when an app is first released it has a simple, discoverable UI. But over time, as people become comfortable w/ the app, its UI becomes less approachable to suit the needs of more expert users and new users are left behind (snapchat for example).

In my opinion a touch screen is fine for consumptive devices where many different UI elements are required and input is light. Creative devices should have a fixed controller like a keyboard, joystick, etc.

That's cool, but I think that most people here agree that mice are a pretty terrible method of interfacing with a computer.

I think mice are excellent, and I think their longevity has proven that people like them, understand well how to use them, and that they are very useful for many computer tasks.

Touchscreens and digitizer tablets and trackballs and all sorts of alternate input methods have existed for thirty years or more, yet they have never broke into the mainstream on desktop computers for anything but niche applications.

The ergonomics and precision of a touchscreen on a desktop or laptop computer are generally abysmal. I can easily move my mouse one pixel at a time when doing graphics work and such with a mouse. Stuff like this is impossible with a touchscreen and big clumsy fingers.

Creative devices should have a fixed controller like a keyboard, joystick, etc.

Or maybe a mouse...

Somebody is a webdev who works in a terminal all day.

Mice are great. They can be used for drawing, dragging, context menus, scrolling, and to quickly and precisely select and manipulate distant controls. Those things are significantly more tedious when performed with a keyboard, or require pre-configured hotkeys. The mouse has stuck around for a reason.

Even worse, I am an operating systems dev. But to be specific, I agree that mice are a useful tool for some tasks. But novices use them like a hammer and everything becomes a nail.

Want to play an FPS game against me? you will use any combination of keyboard, joystick, touch screen you prefer.

Was that the internet equivalent of a "come at me bro"? My point wasn't that mice aren't useful, it was that UIs shouldn't _only_ use mice. Further, touch screens are in trouble because they are just like mice, but they aren't paired with a fixed controller.

Better stock up on mice to get through all the FPS games you have to play during a normal workday.

Jokes on you, I was an Esports "athlete" (World of Tanks)! :P

Keyboards are only faster if you have spent the time learning the shortcuts for a specific application.

With a mouse, you can use any application right away with no learning curve, and also interact with the content itself better than with a keyboard.

That is exactly my point. Touch screens don't have the faster thing to fall back on, so apps end up optimizing the touch interface for speed at the expense of discoverability.

> Every time you use a mouse, you have to re-orient your self to the state of the mouse (its current location and its use within a given application). That is why we spend so much time optimizing applications for use with a fixed controller (a keyboard).

No. This is wrong. The mouse is faster. Apple did that research, too -- and the mouse won every time:

https://asktog.com/TOI/toi06KeyboardVMouse1.html

Recognizing and pointing is always faster than remembering command sequences you have to press/type in. It imposes less of a cognitive burden too, which means you can focus more on the task you're doing and less on the interface itself.

Apple absolutely made the right call with a mouse-centric UI in 1984 and they made the right call with touch in 2007. That's why they're Apple -- you don't become the most successful and influential tech company in the world by making stupid choices.

No, I can definitely hit Alt-Q to close a window, or Ctrl-k to delete a line of text in emacs, before I even finish reaching for the mouse. I'll buy that some tasks maybe faster with a mouse, but not everything.

I remember Wozniak once said that if you wait long enough, what you said will be true.

Apparently he said once (a looooooong time ago) that a mouse is a useless device and nobody would need one. So with the use of the touch-screens mouse use will be reduced.

I have to say that although my laptop does have a very good touch-screen, I still use my mouse for 80% of the actions. Perhaps when browsing or changing windows (to continue browing) I may be using my fingers. But try to select text and change it to Bold on Word, when on the highest possible screen resolution, and well.. you know what happens.

Of course for small portable devices, such as a phone, to write a small note it 'good enough' when using the software keyboard, and the fact that the usual mode on smartphones is 'one-app-at-a-time' displayed on the screen, makes for a simpler UI and UX.

> Put any user interface up. And a pointing device. We solved it with the mouse. We solved this problem.

They solved the problem... by putting "exit, maximize, minimize" buttons at the top-right (or top-left) of every window. By putting "file, view, edit" dropdowns at the top-left of ever application. By putting "back, forward, refresh" at the top of the browser.

Not every application- I don’t see those buttons when dialing a phone number, or when watching a video or playing a game.

...and also by still having a keyboard with a fixed and standardized layout, which Jobs' argumentation conveniently ignores. I'd like to see how much of a success the Macintosh would have been with just a mouse ;)

The problem of how to use a fixed set of buttons for a variety of purposes was solved much longer ago by the designers of (military) aeronautical instruments: you put them on the edge of the screen which displays labels next to currently active buttons. You designate a few buttons to have fixed functions - that would be the home, back and switch buttons on Android - and leave the rest to be used by the current application.

As used by ATMs. Horrible UX in the age of touch screens. Why is this screen not responding to my touch? Oh, buttons...

And even then they don't align to the on-screen text.

Nothing keeps you from making those labels touch-sensitive 'in the age of touch screens'. That does mean they need to be bigger than necessary and it has the usual problems associated with touch screens - you need to visually locate the touch zone as there is no true haptic feedback - but it is possible.

There is no reason for the screen text not to align to the buttons other than sloppy design.

Like old Nokias. The ones I'm familiar with had two unlabelled buttons at the top of the keypad, with on-screen labels directly above them to signify what they did.

Yes, the system has been used in a limited fashion (two or three buttons on the bottom of the screen) by some manufacturers, Siemens (and their DECT-offspring Gigaset) is another example, as is Philips.

If your purpose is to allow the device to be used when it doesn't have your full attention - like a car radio, for example - I don't think that's a good solution.

I have a phone with a knurled power button, and a smooth flashlight button next to it. It's a minor example, but dynamically remapping those buttons would destroy a lot of their utility for me.

That is false analogy as you're opposing fixed-function devices - lamp and power - to a device which in principle has an unlimited number of functions. Even if it could also contains an unlimited number of buttons this would still not make it usable in situations where the user has to divide his attention between the device and some other task of higher importance like flying the aircraft and keeping sight of the opponent. That is just why they came up with this interface since it combines the advantages of haptic memory - button three on the left side does this in that situation - with unlimited functionality. All that it takes is a sure way to get the device in 'that situation' and a modicum of experience. A touch screen will never have this possibility as the user always needs to look at the screen to guide the digit to the target - unless the whole screen is turned into a single button of course.

Buttons on the side of the screen are not the correct user interface for all activities, as an example they would be inconvenient for browsing the web [1] and useless to sign for delivery of a parcel.

[1] ...but not impossible: number the links on the screen as if they are footnotes, number the buttons on the side, click the button which lies next to the desired number. If there are more links than there are buttons the last button can be used to select the next set of numbers. If there are zillions of links the last two buttons can be used to page through sets of numbers. Inconvenient, maybe, but it would work.

Steve Jobs was only correct in the context of a media consumption device. Having only physical buttons would prevent apps like YouTube from allowing users to select a timeline position to jump to that position in a video clip, and touchscreens are certainly better hardware for games like Bejewelled.

Where he was wrong is that professional use requires muscle memory to build mastery of the tool, and that the flexible nature of touchscreens fundamentally prevent people from developing muscle memory. Professionals derive enormous benefit from fixed-button interfaces that are impediments or limitations for the mass consumer market.

(Automobiles are the main exception to this rule, since they require blind usability, but most consumer products don't require blind usability.)

That's great, except it's not even remotely relevant to kitchen appliances.

Those have existed for decades with the same functions and each one of them does a particular task very well. Nor do you do anything on those devices that involves reading a lot of text or watching full screen video where buttons (or UI in general) could get in the way - the polar opposite of smartphones.

But what if you need to tweet using your smartfridge??

But seriously, you are exactly right. There is no need to retrain the user for no benefit.

I think that we've gotten some fantastic UI/UX designs that came from having to work within the limitation of pre-existing controller inputs. You also have the benefit of being able to assume that the user has some level of competency with the controller. Even with Jobs' example, UI/UX on computers generally remains the same within a given ecosystem as changing them forces your users to re-learn things that they already knew how to do. Individual products on that system mostly congregate to using the same patterns too so that they can reap the benefits of users being pre-trained to look for things to perform specific actions.

>And what happens if you think of a great idea six months from now? You can’t run around and add a button to these things.

No, you create a new product with the new buttons and sell something different to your competitors, now all the phones look the same, on top of that you have only 2 Operating systems to choose, there used to be creativity in the design of devices now they cycle from round edges to square and back.

But even iPhone have fixed hardware buttons. Why they are not replaced by pixels on the screen? Can you cite?

They replaced the home button with a digital indicator. You need a physical power button that'll work with the screen off so you can turn the phone on. Physical volume controls allow you to change the volume in your pocket without looking at your screen, and also allow you to force restart the phone. Not sure why the mute switch is still physical, but I don't mind it being physical (although it does get dirty over time).

So, if we need to do something without looking at the screen, we should keep hardware buttons, right? And if we don't need that, if we are looking at the screen _with_ buttons already, then hardware buttons are redundant.

Except in the case where the stupid thing freezes and you can't interact with it anymore, then you need a physical button to smother it so you can start it back up again.