Professional Documents
Culture Documents
3 - NormanNielsen
3 - NormanNielsen
Gestural Interfaces:
A Step Backward In Usability
Donald A. Norman
Nielsen Norman Group and Northwestern University | norman@nngroup.com
Jakob Nielsen
Nielsen Norman Group | nielsen@nngroup.com
One step forward, two steps back. screw things up. Nielsen put it But the place for such experi-
Once again, the usability crisis this way: “The first crop of iPad mentation is in the lab. After
is upon us. We suspect most of apps revived memories of Web all, most new ideas fail, and the
you thought it was over. After designs from 1993, when Mosaic more radically they depart from
all, HCI certainly understands first introduced the image map previous best practices, the more
how to make things usable, so that made it possible for any likely they are to fail. Sometimes
the emphasis has shifted to more part of any picture to become a a radical idea turns out to be a
engaging topics, such as new UI element. As a result, graphic brilliant radical breakthrough.
applications, new technological designers went wild: Anything Those designs should indeed
developments, and the challeng- they could draw could be a UI, ship, but it’s important to real-
[1] Norman, D. A. es of social networks and ubiq- whether it made sense or not. ize radical breakthroughs are
“Natural User Interfaces
Are Not Natural.” inter- uitous connection and commu- It’s the same with iPad apps: extremely rare in any discipline.
actions 17, 3, (2010); nication. Well, you were wrong. Anything you can show and Most progress is made through
http://interactions.acm.
org/content/?p=1355/ In a recent column for interac- touch can be a UI on this device. small and sustained incremental
tions, Norman pointed out that There are no standards and no steps. Bold explorations should
in the rush to develop gestural expectations.” [2] remain inside the company and
(or “natural”) interfaces, well- Why are we having trouble? university research laboratories
tested and understood standards Several reasons: and not be inflicted on any cus-
of interaction design were being • The lack of established guide- tomers until those recruited to
overthrown, ignored, and violated lines for gestural control participate in user research have
[1]. Yes, new technologies require • The misguided insistence validated the approach.
new methods, but the refusal to by companies (e.g., Apple and There are several fundamental
follow well-established principles Google) to ignore established principles of interaction design
[2] Nielsen, J. “iPad leads to usability disaster. conventions and establish ill- that are completely independent
Usability: First Findings Recently, Raluca Budui and conceived new ones. of technology:
i n t e r a c t i o n s S e p t e m b e r + O c t o b e r 2 0 10
46
OPINION THE WAY I SEE IT
• Reliability: Operations should permanent controls at the bot- Android, the permanently vis-
work. Period. And events should tom of the screen: back, menu, ible back button provides one
not happen randomly. home, and search. They are method, but sometimes the task
All of these principles are rap- always visible, suggesting they is accomplished by sliding the
idly disappearing from designers’ are always operative. True for screen to the right. The back but-
tool kits, aided, we must empha- three out of the four—not for the ton has a major flaw, however.
size, by the weird design guide- menu button. This visible menu Push the back button to go to the
lines issued by Apple, Google, and button implies there is a menu previous page, then again, and
Microsoft. available, but no, many applica- then again: Oops, suddenly you
What are we talking about? Let tions (and places within applica- are out of the application, never
us explain. tions) don’t have menus, and having been warned that the
even those that do don’t always next button-push exits instead
Visibility have them everywhere. There is of simply going back. (The same
Nonexistent signifiers. To delete no way to tell without pushing flaw exists on the Blackberry.)
an unread message in Mail on the button and discovering that The back button moves the user
the iPhone, swipe right across nothing happens. (Actually, it through the “activity stack,”
the unopened mail and a dialog means multiple pushes because which always includes the origi-
appears, allowing you to delete the lack of a response the first nating activity: home.
the item. Open the email and the time may reflect the unreliability This programming decision
same operation has no result. In of the technology.) should not be allowed to affect
the Calendar, the operation does Worse, when on the home the user experience: Falling off
not work. How is anyone to know, screen, pushing the menu will the cliff of the application to the
first, that this magical gesture occasionally bring up the on- home screen is not good usability
exists, and second, in which set- screen keyboard. Usually a practice. (Note too that the stack
tings it operates? second push of the same key on the Android does not include
With the Android, pressing and undoes the action done by the all the elements the user model
holding on an unopened email first, but sometimes, the sec- would include: It explicitly leaves
brings up a menu that allows, ond push brings up a menu out views, windows, menus, and
among other items, deletion. that floats above the material dialogs.) Yes, provide a back but-
Open the email and the same being displayed (The keyboard ton—or perhaps call it a dismiss
operation has no result. In the does not always appear. Despite button—but make it follow the
Google calendar, the same opera- much experimentation, we user’s model of “going back,” not
tion has no result. How is anyone are unable to come up with the programmer’s model that
to know, first, that this magical the rules that govern when is incorporated into the activity
gesture exists, and second, in this will or will not occur.) stack of the operating system.
which settings it operates? Among other things, it should
Whenever we discuss these Feedback have a hard stop at the top level
examples with others, we invari- Both Apple and Google recom- of the application. The forced exit
i n t e r a c t i o n s S e p t e m b e r + O c t o b e r 2 0 10
ably get two reactions. One is mend multiple ways to return to from the application is wrong.
“Gee, I didn’t know that.” The a previous screen. Unfortunately,
other is, “Did you know that if for any given implementation, Consistency and Standards
you do this [followed by some the method seems to depend Whatever happened to the dis-
exotic swipe, multifingered tap, upon the designer’s whim. tinction between radio buttons
or prolonged touch] that [the fol- Sometimes one can swipe the and checkboxes? Radio buttons
lowing] happens?” Usually it is screen to the right or downward. meant selection of only one out
then our turn to look surprised Generally, one uses the back of all the possibilities: Selecting
and say, “No, we didn’t know button. On the iPhone, if you are one precluded the selection of
that.” This is no way to have peo- lucky, there is a labeled button. others. Check boxes, however,
ple learn how to use a system. (If not, try swiping in all direc- allow one to select multiple alter-
Misleading signifiers. For tions and pushing everything natives. Now, with these new sys-
Android phones, there are four visible on the screen.) With the tems, check boxes can work any
47
Design and Usability
When users
think they did one
thing but actually did Discoverability
The true advantage of the
surface, or tilted at an angle? All
varieties now exist.
Graphical User Interface (GUI) Sensitive screens give many
something else, they was that commands no longer opportunities for accidental
had to be memorized. Instead, selection and the triggering of
detailed interface guidelines for screens will range from tiny to easy to commit the error. When
their products. Unfortunately, huge, conceivably wall-size (or at users think they did one thing
the guidelines differ from one least, whiteboard-size). However, but actually did something else,
another, in part because differ- gestures that work well for small they lose their sense of control-
ent companies wish to protect screens fail for large ones, and ling the system because they
their intellectual property by not vice versa. Small little checkbox- don’t understand the connection
allowing other companies to fol- es and other targets that work between actions and results. The
low their methods. But whatever well with mouse and stylus are user experience feels random
the reason, proprietary stan- inappropriate for fingers. Larger and definitely not empowering.
dards make life more difficult for screens have their own problems Some reliability issues can be
everyone. For sure, they under- with control sizes. Are the new alleviated by following usabil-
mine the main way in which controls to be used while held ity guidelines such as using
users learn from each other. in the hand, laid flat upon a larger objects and surrounding
48
OPINION THE WAY I SEE IT
them with generous click zones. the fingers, providing some visu- The new displays promise to
Others are inherent in any new al excitement, but we still need revolutionize media: News and
technology that will have its to work out the display dynam- opinion pieces can be dynamic,
bugs—that much more reason ics, allowing the items to gather with short video instead of still
to enhance user empowerment speed, to keep going through photographs and adjustable
by designing according to the a form of “momentum,” yet to figures that can be manipu-
other interaction principles make it possible to see where one lated instead of static diagrams.
we have listed in this article. is in the list while it whizzes past, Consumer Reports could publish
Lack of undo. Undo! One of and to enable rapid stopping once its rating tables with reader-
the most brilliant inventions the desired location seems near. controlled weights, so each
of usable computer interfaces Although pinching and spread- viewer would have a tailored set
seems mostly to have been ing seem like natural ways of of recommendations based upon
forgotten. It is very difficult to zooming an object out and in, standardized test results.
recover from accidental selec- when the dynamics are badly The new devices are also fun
tions or checking of boxes. First, set, the movements are difficult to use: Gestures add a welcome
the result often takes you to a to control. Different applica- feeling of activity to the other-
new location. Second, it may tions today use different rules, wise joyless ones of pointing and
not even be obvious what action which end up confusing people. clicking.
got you there. For example, if Moreover, even if they could, But the lack of consistency
a finger accidentally scrapes not all places allow this: another and inability to discover opera-
an active region, triggering an source of confusion. tions, coupled with the ease of
action, there is almost no way to Rotation and tilting the device accidentally triggering actions
know why the resulting action are also often used to change from which there is no recovery,
took place because the trigger the display, although for some threatens the viability of these
was unintentional. applications, such as reading, systems.
it has been found necessary We urgently need to return to
Novel Interaction Methods to provide a lock to prevent our basics, developing usability
Gestural systems do require the otherwise natural rotation guidelines for these systems that
novel interaction methods. of the displayed image that are based upon solid principles
Indeed, this is one of their vir- would prevent easy reading. of interaction design, not on the
tues. We can tilt and shake, whims of the company-interface
rotate and touch, poke and The Promise of guidelines and arbitrary ideas of
probe. The results can be Gestural Interfaces developers.
extremely effective while also The new interfaces can be a
conveying a sense of fun and pleasure to use and a pleasure to About the Authors Don Norman
and Jakob Nielsen are co-founders of the
pleasure. But these interaction see. They also offer the possibil-
Nielsen Norman group. Norman is a profes-
styles are still in their infancy, so ity of scaling back the sometimes sor at Northwestern University, visiting pro-
it is only natural to expect that heavy-handed visual language fessor at KAIST (South Korea), and author.
i n t e r a c t i o n s S e p t e m b e r + O c t o b e r 2 0 10
a great deal of exploration and of traditional GUIs that were His latest book is Living with Complexity.
Norman can be found at jnd.org.
study still needs to be done. designed back when nobody had
Shaking has become a stan- seen a scrollbar. In the early Nielsen founded the “dis-
dard way of requesting another 1980s, usability demanded GUI count usability engineer-
ing” movement for interface
choice, a choice that seems to elements that fairly screamed
design and has invented
have been discovered acciden- “click me.” several usability methods,
tally, but that also feels natural. Desktop GUIs are already less including heuristic evalua-
Note, however, that although it neon than Windows 3.0, and we tion. He holds 79 U.S. patents, mainly on
ways of making the Internet easier to use.
is easy and fun to shake a small can afford to dial back the visual
Nielsen can be reached at useit.com.
cell phone, shaking a large pad prominence a bit more on tab-
is neither easy nor much fun. lets, which will further enhance
Scrolling through long lists can their aesthetics. But dialed back
DOI: 10.1145/1836216.1836228
now be done by rapid swiping of doesn’t mean invisible. © 2010 ACM 1072-5220/10/0900 $10.00
49