A Science of Practology

In the past, shoes could stink

In the present, shoes can blink

In the future, shoes will think

- sign at the MIT Things That Think lab

            Every so often, when I try to boil some water on the stove I turn on the wrong burner. 

            A minor mishap to be sure.  Until you consider that I have tried this thousands of times in the twenty years that I have owned this one stove.  I would feel upset except that everyone I talk to seems to have this same experience.  The difference is that most people end up blaming themselves, annoyed at not being able to learn something so simple.  I am annoyed too but at the stove makers rather than myself.  This sense of superiority is not due to an overheated ego.  It is the result of my work in gizmotics, especially as a designer computer games.

            You may not see any relationship between boiling a pot of water and clicking on an icon but there is an important one.  The connection is in that slippery area of contact between machines and humans known as the interface.

            Interface refers to the parts of machines that we humans use and manipulate.  In most cases, this includes some kind of control panel.  When they work well – an old dial radio for example – there is a natural connection between the function of the device and our behavior.  Dials are good controls because they are analogous.  Turn more, get more.              But dials are a dying breed.

            In the evolution of interfaces, levers begat switches, which begat dials, then buttons, which have now morphed into clickable icons.  These ghostly objects are increasingly going to be the way we manipulate the world.  As we continue our sprint from the industrial to the virtual, the interface problem becomes tricky because there is no direct link between the action and the result.  This makes it even harder to manage a standard problem of interface design...making the controls easy to use rather than simply efficient. 

            The controls on a new cable router, to take one example among many, were designed to accommodate the mechanism of the box, not the layout of my hands and intellect.  There were four tiny buttons around the rim that had to be pressed in bizarre sequences to access dozens of functions, all while looking away at the TV screen.  It came with a thick instruction manual that would have adding another layer of holding and looking.  Great plan for a smart octopus, not so great for me.

            This same focus on engineering over manipulation is the reason most people never use most of the functions on most of their devices.  The buttons are made to accommodate the layout of the electronics, not the patterns by which people actually do things.

            The stove is a perfect example, particularly on older models like mine where the burners are set up in a square array while the knobs are arranged in a line.  Without a visual relationship, there is no way to relate the two.  It was designed for efficient function…a square of burners and a row of dials fit best in the limited space.  

            However, if my use of the device were the key factor, the patterns of knobs and burners would match, or be color coded, or any number of other alternatives.  And my phone would not have a 40-page instruction booklet outlining dozens of arcane features I will never use.  It would help me do one thing….talk to someone. 

            One page.

            What we need to make sure that interfaces work is a science of Practology.

            A way of thinking about machines that focuses on how things are actually going to be used.  After all, the interface is essentially what protects us from the innerface, the inner workings of the device.  Good interfaces should be intuitive, simple, and immediately useable by both novices and sophisticates.  Like a piano keyboard.  One note, one sound, left to right up the scales.  Chopsticks or Rachmaninoff is up to you. 

            The touch screen at my ATM has a great interface: simple options, simple actions, and no moving parts.  The choices are grouped in ways that allow me to make a limited number of decisions at a time.  Very nice.

            On the other hand, bad interfaces force us to relearn the procedures every single time.  Like some software applications or Web sites that hit you with a dizzying swarm of icons all equally available so that you can never quite organize your own choices.

            Some interface designs only become comfortable over time as our habits rise to the occasion.  It is hard to imagine now, but there is no particular reason for directing a car with our hands on a wheel while adjusting the speed with foot pedals.  Early cars had tillers not steering wheels and the brake was manually operated.  The mass produced version could just as well have reversed the controls we have now: pedals for direction – press left go left, press right go right; – and handles for speed – push forward to accelerate, pull back to slow.  Yet now, 80 years after the pattern was established, wheels and pedals seem like the most natural setup in the world.  Ditto for the QWERTY layout on a keyboard.

            So why can't I turn on the right burner on the stove as easily as I can gun the gas or type the word “gas”?  Because of repetition.  In spite of trying a few times a week, I really only practice on my stove intermittently while I am continually rehearse the wheel and pedal all the time I drive and the keys as I type this.  A professional water boiler would probably not have my problem.

            Designers can therefore get away with lousy interfaces if people use them all the time and repeat the awkward behaviors over and over.  That is why you can become familiar even with a befuddling word processor or email system.  But if the use is intermittent, there is no excuse...it is the fault of the design. 

            No practology at work. 

            And either way, habit or utility, it is the designer's job to take our needs into account, not the other way around.