On Design

I’ve just finished reading a book on user interface design principles – The Design of Everyday Things by Dr. Donald Norman. A very famous book by a very famous author in this field (cognitive sciences), written in the early 1990s.

It does have some very good points, but I won’t particularly recommend it unless you are into this kind of things, because the writing style is somewhat plain, and very repetitive.

Here is my attempt at a 30 seconds summary (most of the examples are straight from the book) –

What makes a good design?
Visibility – controls have to be visible. For example, give closets a handle bar.

Make it “natural” for the user to do the right thing – for example, on the side of the door where you want the user to push, give it a large panel. On the side you want the user to pull, give it a handle bar. No “push”/”pull” labeling needed.

Natural mapping – the arrangement of the controls, and what they do, should correspond to the “mental image” the user have of the system. That’s what makes some interfaces more intuitive than others. Avoid arbitrary mappings. For example, if you have 4 knobs that control 4 stoves, arrange them in the same pattern as the 4 stoves. This way the user can easily remember which knob controls which stove, and no labeling is needed. Arrange light switches to mimic how the lights are themselves physically arranged. On the other hand, if you have 4 switches in your house that controls 4 lights, and 1 that controls a current through your cute kitten, don’t use 5 identical switches arranged with equal spacing. People will group them together mentally, and when they see the first 4 all turn on a light, they will try the last one.

Assume the user will make mistakes – make most mistakes cheaply (in terms of effort/time) reversible, and irreversible ones very hard to make. Don’t allow flicking a switch to crash an airplane.

Provide immediate feedback – let the user know their action has taken effect. Silence is not golden. Don’t put the light switch of a room in another room. Don’t tell the user to press an arbitrary sequence of keys, and only give feedback at the end. Don’t put a “turn the light on after 5 minutes” button. If you do, at least add an LED that turns on immediately.

Exploit cultural constraints – Left knob controls hot water, right knob controls cold water. Don’t reverse them for no reason. And don’t arrange them up and down. This is knowledge in the culture, everyone already has it. Don’t make a knob look like a push button, vv.

That’s all I remember off the top of my head.

Now, let’s look at an example of a terrible design.

Apple’s spanking new Magic Mouse –

They took out all the buttons, and replaced it with a big multi-touch surface (and call it the best invention in 21st century, the thing that will change our lives forever, and give our lives new meanings, yada yada, etc).

How many good design principles did it violate? Let’s see.

1) No visibility. The user doesn’t know where to click. There are no clearly (or ANY) distinguished buttons, or scroll wheel. The whole surface is uniform. Best example of flushing usability down the drain for aesthetics. You can also not find the buttons by touch.

2) No feedback. When you click it, you won’t know you clicked it. There is no haptic (touch) feedback. You have to guess.

3) Arbitrary mapping. There is no way to tell which way points forward (this is from reviews). To do a middle button click (which is admittedly not so common in Mac, but very common in UNIX/Linux), you need to hold down the ctrl key (where did that come from?).

It’s also uncomfortable to hold, and scrolling occasionally triggers clicks (this is from reviews).

Yes, it’s truly a new invention. No one has done it before. Just like how no one has tried to build a house out of poop. It doesn’t work. People haven’t done many things not because they never thought of it, but because it doesn’t make sense.

What do you get from all this uncomfortable compromises? Exactly nothing. Horizontal scrolling? Many mice have that already. Nothing new.

People have used buttons on mice for decades, without much change. That says quite a lot about the design.

A button is cheap, findable by touch, gives ample feedback, and is durable. What more can you ask for?

They are just changing for the sake of changing. They couldn’t think of a way to improve it, so they go the other way. That’s a whole lot easier.

I imagine it will be a negative example in design textbooks for years to come.

By the way, the iPhone 4 antenna incident. Did you see how dirty it was at their press conference? They tried to shift attention to a totally unrelated issue (flesh blocking RF because of high water content – everyone knows that), when the REAL issue is how fingers can easily bridge the 2 antennas, as well as ruining the impedance matching. If the issue is really just flesh blocking RF, bumpers wouldn’t have helped at all. Then they went into this everyone-lies-about-signal-strength-and-now-we-are-correcting-it business, when THEY were the ones lying about signal strength. Seriously, there are more honourable ways of doing business.

I love how half of my posts deteriorate into Apple-rant. I shall stop now. They are not worth my time.

4 thoughts on “On Design”

  1. Thank you, I was waiting for someone to agree with me this whole time. It really annoys me when my friend says how Apple interfaces are intuitive. How the hell am I supposed to know that you’re supposed to double swipe left with two fingers?

  2. I like old keyboards they give good tough feedback. Helps me type without looking at the keyboard, something that I cannot do with new ones!

    PS it seems like apple products are more of a fashion statement than helpful equipment these days.

    1. Yeap! I am still using a spring loaded keyboard from the early 1990s (no exaggeration. It’s almost as old as me).

Comments are closed.