The Hand-Dryer's Secret Handshake26 May 2008
“Why did it do that?” and “what is it doing now?” are, expletives deleted, probably two of the questions most frequently asked by confused computer users. As experienced users, the ability of software developers to reason about what a desktop application or website is doing means that they often have difficulty empathizing with users who don’t have the same stock of experience. It’s not the developer’s fault, it’s a natural consequence of being human: we learn from experience. But once learned, knowledge is difficult to forget, temporarily, when we need to look on a design with inexperienced eyes.
Familiarity with the way software operates internally helps experienced users understand and navigate poorly-designed user interfaces. This familiarity is often so ingrained that experienced users don’t see such interfaces as poorly designed. Any problems experienced users do have are often worked around and go without comment because, well, that’s just the way computers are. In The Humane Interface, Jef Raskin describes the operating system as the thing users have to hassle with before they have to hassle with the applications. Maybe we should call the operating system and its applications hassleware. One of the problems we have as interaction designers is making software developers understand firstly, that there is a problem with the hassleware they produce, and secondly that it is vital that we all do something about it.
When we encounter a new object, we build a mental model of how we think that object works. The model of how a piece of software works, called the implementation model, should be masked and simplified for users by the representation model, which is embodied in the user interface. Our goal as interaction designers is to produce a representation model that matches the user’s mental model as closely as possible. Because developers build their mental model from the implementation of the software rather than its user interface, it’s hard for developers to see the flaws in their interfaces.
Developers cannot readily empathize with users because users create their mental model solely from the user interface. In About Face, Alan Cooper’s solution is to remove the need for developers to empathize with users altogether: the people that design the interaction shouldn’t be the ones that implement it; empathy is the sole domain of the interaction designer. For the time being at least, user interfaces will be designed and implemented by the same people. This group of developers needs our help. Providing developers with interaction design training is not enough to help them develop more usable software. We need to help developers empathize with users.
One way to engender empathy with users is to show them examples of poorly-designed non-computer interfaces. The Bad Designs website, for example, provides many examples of poorly-designed hardware. The advantage of highlighting the usability problems of non-computer interfaces is that software developers are less able to guess the reason for unusual or unexpected behavior. Without their insider knowledge, software developers are more vulnerable to usability problems. This vulnerability puts them on the road to user empathy.
The electric hot air hand dryer is a good example of an everyday non-computer interface because it seems too simple to have any usability problems. I’ve encountered two types of dryer designs and both have annoying usability problems. The first design is probably the most common: the push-button dryer. To get hot air you push the button. The dryer produces hot air for a short period of time and then shuts itself off. If you want more air, you push the button again. Simple.
The second dryer design has a much simpler interface. There is no button to press; you put your hands under the air vent to activate a sensor that turns on the dryer. The dryer produces hot air until you remove your hands. Even simpler. Alan Cooper refers to extra unnecessary operations in user interfaces as excise. Removing the excise of the button simplifies the interface of the sensor-activated dryer and makes it easier to manufacture and maintain.
The problem with some sensor-activated dryers is that the hand position required to activate the sensor is further back than the air vent. Because these two positions are not aligned, placing hands under the dryer ready for hot air does not necessarily switch on the dryer. The hands must be moved backwards to activate the sensor and then forwards to receive the hot air. A panel often provides instructions but, as Donald Norman notes in The Design of Everyday Things, if a simple device needs instructions, the design of that device is faulty. If users can be bothered to read them, the instructions only confirm what the user is already doing: Place hands under dryer for hot air. If only there was a button to press… So much for the simpler interface.
This problem is exacerbated by sensor-activated dryers that, incorrectly, follow the push-button dryer’s model of producing air for short periods. Because the hands are under the vent rather than the sensor, the dryer switches off while the hands are still under the dryer. The hands must be moved backwards to activate the sensor and forwards to receive hot air. After using a number of sensor-activated dryers, one can learn the hand gesture necessary to coax them into action. I call this the hand-dryer’s secret handshake.
The secret handshake is a poor design because the dryer should dispense air while hands are under the dryer. Maybe there is a safety concern that users will burn their hands if exposed to too much hot air. However, I think this concern is unwarranted because of the universal, culture-free rule that we learn at a very young age and never forget: hot things hurt; if the air from the dryer is too hot, pull your hands away.
Even after learning the correct hand gesture to start the dryer, users can still be stymied by an insidious design decision that affects both push-button and sensor-activated dryers. After dispensing air, some dryers will not dispense air again when the button is pushed or the hand gesture is made until a certain period of time has elapsed. My accurate mental model of how dryers work—push button or put hands under dryer to get air—has just taken a nosedive. My mental model now has to somehow explain why, if I need more hot air, performing the same action to restart the dryer that I used to start it won’t work until the dryer feels like it. Maybe the dryer has just broken. Maybe it has a thermostat and has become too hot (aren’t dryers supposed to get hot?). Anyway, all I know is that I can’t get the dryer going again. If the dryer was a software application, my experience might help me. But I’m at a loss because I experience hand dryers only through their user interface; I have no implementation model to help me out. Every time I use a poorly-designed hand dryer, or any other non-computer user interface, my inability to explain what is going on reinforces my sense of empathy for inexperienced computer users.