Steven Pemberton : Views and Feelings

September 1998

Our Subliminal Art

The shortest zen story I know of goes like this:

"What's this water the other fish talk about, Mummy?"

As far as I am concerned, it perfectly reflects what the user interface should be about: the user shouldn't even be aware of the interface because they are completely immersed in the task, and not drowning in the details of the user interface.

However, this carries a danger for us who work in the HCI field. It was well described by someone to me recently: "When you make user interfaces that are as easy to use as coffee machines, people treat you like a plumber". (Though on reflection he ought to come and see some of the coffee machines at my work).

Many lay-people think that HCI is all surface: that it is somehow only about menus, buttons, and clicking. But as we know, HCI is much more than that. In fact the surface is the least interesting part. Apart from surface (how you do it), there is also function (what you are able to do), and maybe the most important part, what I can best describe as 'attitude'.

Attitude is the zen of interaction: it is hard to learn, very hard to teach. It is almost as if you have either got it, or you haven't. Which is also true of zen: the old zen masters would teach their pupils by analogy, and if that didn't work by hitting them over the head with their stick.

The best way to hit an interface designer over the head with a virtual stick is to let them sit next to someone trying to use their system (and I wish some of them would come and sit next to me, so that *I* could hit them over the head with the stick). The idea is that having seen that they don't yet understand what makes a good interface they go away and think, and eventually achieve enlightenment.

Interestingly, attitude can exist independently of the surface. There are products with the right attitude and the wrong surface. Even before HCI had a name, there were products with the right attitude. Let me try and shock you by naming some unlikely candidates. Algol 68 was the first computer artifact where I noticed this attitude: the language has a bad image because of its bad documentation, but it was the first time that I ever heard people talking about consistency in the way we mean it. Unix had an underlying attitude that in 1975 was shocking in its simplicity. Emacs is another example.

I have chosen exactly these examples because I want to separate surface from attitude. All these examples have been used at one time or another as examples of bad interfaces, and yet despite this they continue to have enthusiastic, even fanatically enthusiastic, users. The applications can't be doing everything wrong! I would contend that the users of these systems having broken through the learning barrier of the bad surface, so much appreciate the underlying attitude that they continue to use them with pleasure.

I'm not arguing that applications should be like Emacs. I'm arguing that without the right attitude, no amount of buttons and menus and colour will turn an application into a good one. To try and use a modern example, think of the first time you used Quicken, and compare that with any current email program. Email programs haven't yet got it.

Pass me my stick.

First published in ACM/Interactions, September 1998

Other Posts