Brenda Laurel, Editor
Addison-Wesley Publishing Co., Reading, MA, 1990
523 pages, $26.95
The Art of Human-Computer Interface Design
The Art of Human-Computer Interface Design is an anthology of essays, musings, interviews, and editorials on the topic of human-computer interaction. This book pulls together thoughts from a wide range of disciplines, ranging from cognitive psychology to ergonomics to sociology to cybernetics. The writers include some renowned thinkers in the field of human-computer interaction.
How This Book Came to Be
An explanation of the genesis of this book gives a valuable background as to the intentions and purpose of the book. The idea for this book first sprang from the mind of Joy Mountford, the manager of Apple's Human Interface Group. The Human Interface Group is responsible for establishing and enforcing Apple's Human Interface Guidelines. These guidelines are supposed to help software developers to create new software that maintains a consistent interface with all other Macintosh programs.
The reason Apple likes to capitalize the name of these guidelines is to stress that these guidelines are ignored at great peril to the individual software developer. In other words, the Human Interface Guidelines carry about the same weight as the familiar guideline to eat, sleep, and breathe, if you happen to be interested in good health.
So the book first started out being intended as a training manual for Apple employees. But the book grew beyond its original design: "The project quickly began to grow into an idea for a trade book that would give Apple authors a chance to publish some of their work on interface design." Many of the writings in this book are by people in Apple's Advanced Technology Group, Apple's in house design- team/think-tank. To round out the perspective, a healthy selection of writings by non-Apple employees was thrown in for good measure.
The Difficulty of Summarizing an Anthology
In reviewing an anthology, its hard to sum up the exact scope of the book in several neat phrases. An anthology is by definition a collection of diverse ideas: a smorgasbord of notions. It may seem a tad unconventional to open this book review with a listing of the table of contents, but the table of contents gives a good overview of what this book is about. It's worth listing the entire table here, to give you a feel for the mood and scope of the book. As you'll see below, the book is organized into five broad sections: Creativity and Design, Users and Contexts, Sermons, Technique and Technology, and New Directions.<
Introduction
CREATIVITY AND DESIGN
Introduction
An Interview with Don Norman
Interface and Evolution of Pidgins: Creative Design
for the Analytically Inclined
Tools and Techniques for Creative Design
Interdisciplinary Collaboration
Two Disciplines in Search of an Interface: Reflections on
a Design Problem
Designing the Whole-Product User Interface
Working with Interface Metaphors
Consistency
Prototyping
Some Techniques for Observing Users
USERS AND CONTEXTS
Introduction
Koko's Mac II
Lessons from Computer Game Design
Interfaces for Learning: What Do Good Teachers Know That We Don't?
Lessons Learned from Kids: One Developer's Point of View
How I Learned to Stop Worrying and Love HyperCard: A First Person
Account of a Paradigm Shift
A Writer's Desktop
Building User-centered On-line Help
Managing the Mundane
An Environment for Collaboration
Groupware and Cooperative Work: Problems and Perspectives
SERMONS
Introduction
User Interface: A Personal View
Why Interfaces Don't Work
User Interface Races: Sporting Competition for Power Users
The Evolution of Thinking Tools
The Interpersonal, Interactive, Interdimensional Interface
The Right Way to Think About Software Design
The Noticeable Difference
TECHNIQUE AND TECHNOLOGY
Introduction
Animation at the Interface
New Uses for Color
Recognizing the Symptoms of Hypertext... and What to Do About It
Adventures with Hybrid Systems: Integrating the Macintosh Interface
with External Devices
A Design for Supporting New Input Devices
Gestures in Human-Computer Communication
Talking and Listening to Computers
Illusion in the Interface
NEW DIRECTIONS
Introduction
Hospital Corners
Interface Agents
Guides: Characterizing the Interface
Narrative and the Interface
Conversation as Direct Manipulation: An Iconoclastic View
The "Natural" Language of Interaction: A Perspective on
Nonverbal Dialogues
VIDEOPLACE and the Interface of the Future
Virtual Interface Environments
Through the Looking Glass
What's the Big Deal About Cyberspace?
Integrating Computers and Television
Designing a new Medium
POSTCRIPT: On Visions, Monsters, and Artificial Life
Interface Evolution
REFERENCES
CONTRIBUTORS GALLERY
SUBJECT INDEX
NAME INDEX
If some of the titles of these essays sound a mite bit fanciful, it's because the contents of some of these essays are a mite bit fanciful. Not all the gripping sounding titles are gripping, however. And some of the duller sounding titles have stimulating content.
Lamentably, this review is being written in linear space, and so buttons cannot just be added to those essays worthy of detailed review. So this review will proceed down the above table of contents, commenting on those essays that seem particularly noteworthy.
It's only fair to state one's biases for the reader to take into account. The bias of this particular reviewer is towards the type of writing that is clear and simple enough to be understood by a layperson. And specificity of ideas is of greater interest than vaguely stated ideas. Consequently, any essay mentioning the amorphous phrase, "paradigm shift," immediately has two strikes against it in this reviewer's eyes.
The book starts out with a well-written introduction by editor Brenda Laurel, framing the central issue of the book. Laurel states that an interface can be defined in terms of its functionality. In general terms, an interface is that entity which allows human beings to use a machine in an intuitive way. From this perspective, a doorknob is a highly effective interface for allowing human beings to control the opening, closing, and swinging of doors. A rounded doorknob fits comfortably into a human hand: "the doorknob extends towards the user and its qualities are biased toward the hand."
Do modern day computers serve human needs as neatly and nicely as doorknobs do? Do the computers we use minimize frustrations by allowing us to do our work in the easiest, most intuitive way? Speaking to the issue of frustration, Laurel offers up this gem:
A new version of my favorite word processor offers me twice as many options as its predecessor. Theoretically, I can now customize my environment and achieve more complex and sophisticated goals. But the plethora of options --- and the interface conventions that I must learn in order to deploy them --- leaves me bewildered and tired. Psychologists call this an increased cognitive load. I call it trouble.
Sympathetic bells ring out in the minds of just about everyone who reads the above lines. This is a fine launching pad for a book about human-computer interactions.
The Contents
To delve into each of the essays in this book would take too much space and effort. Far better to examine several essays closely than many essays superficially.
Koko's Mac II: A Preliminary Report
In a somewhat unusual collaboration, Apple Computer, Inc. has joined forces with the Gorilla Foundation to examine the possibilities of language acquisition among higher-order primates. This is not really as far-fetched as it might seem.
Since 1972, Koko the gorilla has been learning a version of American Sign Language (Ameslan). Koko's 600 word vocabulary extends far beyond words for food, bodily function, and commonplace nouns. The aim is to see how far a gorilla's expressive ability can develop in an environment rich in interaction with speaking humans, and rich in other intellectual stimuli. Consequently, a Mac II has been outfitted, along with a 19 inch touch-screen monitor, to serve as a voice for the ideas in Koko's mind.
The design consideration for such a computer speak to the general themes developed elsewhere in this book. Gorilla's don't have the dexterity to use a mouse, but they can work quite happily pointing to icons using a touch screen. Also, a gorilla's frustrations with a computer might end up being manifested in physically hostile ways towards the computer.
So a one-inch thick glass covering protects the screen of the monitor from the, "2,000 pounds of force an excited gorilla can generate." The computer itself is enclosed to protect it from such flying missiles as unripe bananas and gorilla feces.
The software being used in this project includes three different SuperCard stacks. One of the stacks, called KokoPaint, allows Koko to do simple finger painting using the large-sized touch-screen monitor.
The researchers have come up with an ingenious way to help Koko gain a sense that her actions using the touch screen can have effects on the real world. They have set up the computer so that Koko can control the brightness of the lights in her environment by choosing options on the computer screen. To help firm this idea in Koko's mind, the researchers might consider giving Koko a physical dimmer switch to play with too. Such cause-effect toys are bound to fire up neurons in Koko's mind.
This essay about Koko and her Mac II raises some interesting sorts of questions. But what the essay fails to state explicitly is that Koko has developed a fondness for playing the Puzzle desk accessory, under the Apple Menu, after all the researchers have gone home for the night. Koko is also scripting a HyperCard stack to be used to help develop higher-order thinking skills among the various researchers.
Lessons from Computer Game Design
In this essay, Chris Crawford, a computer game designer, treats us to some insights about intrinsically appealing interfaces of arcade-style computer games. He points out that the interface design for an arcade game is far more demanding than the interface design for a productivity program: "The user of a game feels no compulsion to play the game. If the game's interface is clumsy or confusing, the player simply abandons it. Thus, a game's user interface must pass not a relative test but an absolute test, and a harsh one at that. The user interface must be not merely functional, nor even just easy to use --- it must also be fun."
Crawford goes on to discuss three general lessons from game design: 1) Move away from using the keyboard as an input device, 2) Place greater reliance on graphics and sound, and, 3) Emphasize intensity of interaction.
The third of these points goes to the problem of slow computer response time. Modern day users are getting frustrated at having to wait more than a second or two for a computer to respond to a user command. It's not that we are all collectively becoming more irritable. It's just that the wasted seconds of waiting add up to wasted minutes and wasted hours.
The second half of Crawford essay analyzes the problems of nested menus and the increasing complexity we can expect in application programs coming down the pike. These are interesting ideas for the right hemisphere of your brain to contemplate, but don't whisper a word about this increasing complexity to your left hemisphere.
Interfaces for Learning: What Do Good Teachers
Know That We Don't?
Anne Nicol works with Apple's Human Interface Group. Writing from an educator's perspective, Nicol writes about the type of sensitivity that should be an integral part of any design process. One of her messages is that before you understand what works best, you've got to watch carefully where people trip up: "For years, I've been watching children using computers to learn a whole range of subjects and ideas --- from programming to history to creative writing. Children work miracles with the machines and overcome all kinds of obstacles that would stymie their parents and teachers. But what interests me are the places where they stumble or get off-track. I learn from the mistakes children make and the misconceptions they develop as they use computers. And what I learn often has relevance for adults as well as for children, for experts as well as novices."
Following this, Nicol discusses the process of "discovery learning," and what can be done to facilitate the ease with which a person discovers how and why something works. Some of the techniques she suggests include giving cues and overviews for users who are unfamiliar with the territory.
One way to give users a sense of the organization of a program is to carefully choose the placement of options in various menus. A well-constructed menu system enables users to create a mental model of how the program hangs together. Then, using that model, relevant options should be accessible without giving much thought about where to look for them.
Nicol ends her essay by reiterating her educator's point of view: "But long before there is something to take to the children, designers could begin in the classroom by watching good teachers teach. Look for the ways they set up the environment and prepare their students for new concepts. Watch how they coach and guide students; observe the models they put forth for inquiry."
Apple's Human Interface Group is heading in the right direction when it listens to the insights of Anne Nicol.
Lessons Learned from Kids: One Developer's Point of View
Joyce Hakansson is an educational software developer who has created computer activities, interactive videos, and hand-held electronic toys for children. Among her many credits is the design of the computer gallery at Sesame Place. Sesame Place is the child-centered amusement park created by the Children's Television Workshop. (It's located on the outskirts of Philadelphia, not far from Trenton, New Jersey.)
Hakansson opens her essay by remarking that the academic field of child development is not a new one, but that few studies have been done to find out more about the computer's effect on child development. Viewing the world from a slightly different perspective from Anne Nicol, Hakansson has formulated her own informal theories of what works best in child-computer interaction.
One of the lessons she's learned in her dealings with children is that computers can provide an excellent opportunity for children to work cooperatively. "Research studies have shown that in classrooms where students use computers there is more social communication and cooperative problem solving than in other classrooms [Office of Technology Assessment, 1988]. This is exactly the opposite effect predicted by early opponents of technology, who warned of classrooms filled with isolated students staring blankly at video screens."
By working cooperatively in learning activities in elementary school, children can have a first-hand taste of the sweet fruit of collaboration. In some instances, kids will be able to transfer these same social skills into their adult work environments.
Hakansson also examines how point-systems in games can motivate children to challenge themselves at steadily increasing levels of difficulty. What does this have to say about how the human mind reacts to rewards? Maybe application programs for adults should incorporate point systems, so that a word processor, data base, spreadsheet, or telecommunications program would give you extra points if you used the program to its full potential. How come adult computer programs don't have a Hall of Fame, for experts to enter their names into?
Another part of this essay covers the important topic of the subtle ways that young girls are steered away from taking an interest in computers. Part of the fault for this problem lies with software and hardware developers who develop programs with obviously male-oriented themes. Part of the fault lies with newspapers who allow advertisers of such products to claim in bold type: "Every Boy Wants One." (Hakansson reports of such an actual advertisement.) Part of the fault lies in outdated attitudes about sex roles and technology. (i.e. men are responsible for fixing the car, mowing the lawn, repairing appliances around the house, and programming the VCR). And part of the problem is the absence of role models of adult women who take a strong interest in working with computers.
The essay ends up concluding that children can think in novel ways, but they need tools to facilitate and support such thinking. Kids will naturally explore around, and engage in discovery learning, when they are given a suitable environment to do such exploration. Kids are naturally self-motivated, so the goal of software designers is to harness that enthusiasm in a constructive, educational way.
User Interface: A Personal View
Alan Kay develops some interesting ideas in this essay. He starts off by placing computer interface design in a historical perspective: "Of course the practice of user interface design has been around at least since humans invented tools. The unknown designer who first put a haft on a hand axe was trying not just to increase leverage but also to make it an extension of the arm, not just the fist... From the earliest use of interactive computing in the fifties --- mostly for air traffic control and defense --- there have been attempts at user interface design and application of ergonomic principles."
This broader perspective on interfaces in general harks back to Brenda Laurel's introductory remarks about a rounded doorknob being "biased towards the hand." The goal of interface designers, then, is to make the interface biased towards the operations of the human mind.
Kay goes on to talk about how the writings of Marshall McLuhan helped reveal that the computer could serve both as a communications medium as well as a communications tool. For instance, people use computers to help them compose ideas. But they also use desktop publishing programs and presentation graphics programs to help present ideas. HyperCard, and similar programs, are as much a tool to help people create hierarchies of knowledge, as a tool to access those hierarchies of knowledge.
Kay's writing can be a bit dense at times, such as when he compares three "defined mentalities": the enactive, the iconic, and the symbolic. But it's worth following his train of thought, because behind the abstract language are some intriguing ideas.
For instance, in talking about how users are becoming increasingly empowered to choose options in the programs they use, Kay offers up this nugget: "...I believed then (and still do) that the nirvana of personal computing comes when the end users can change the tools and build new ones without having to become professional-level programmers."
Already, there has occurred a blurring between advanced computer users and computer programmers. Given sufficiently powerful software tools, advanced computer users can customize their programs to such a degree as to create something that in early years would be considered a new program. Silently, without much commotion, the advanced computer user has been crossing the threshold into the realm of computer programming.
Kay finishes up his essay developing the concept of agent, or intelligent assistant, that will likely play a part in upcoming computer developments. Computer agents will serve as guides or coaches to help you navigate through information systems. You tell the agent your intellectual interests and biases. The agent will steer you in the direction you'd like to go.
The concept of agency is treated several other times in this book, most fully in the last section of the book titled New Directions.
The Appendices
Any well organized nonfiction book has a good index and a good list of related readings listed by chapter. This book has two indices: one for subjects and one for names. And the list of related readings takes up a full 22 pages at the end of this book. Browsing through the list of related readings turns up books by social theorists, such as Marshall McLuhan, Jean Piaget, and Bertrand Russell, as well as books by computer theorists, such as Marvin Minsky, Seymour Papert, and Ted Nelson.
Truth is, though, that dividing these authors into two camps creates an annoying sort of distinction. All these authors are concerned with how people use tools. Some concentrate on the people side of things. Some concentrate on the tools. Ultimately all are interested in the same thing.
The only dismaying part of the related readings section is that some of the listed references are totally inaccessible to anyone but professional researchers. Who has the time or energy to go and find: Kling, R. "The Social Dimensions of Computerization." Plenary address given at CHI+GI '87 Human Factors in Computing Systems, Toronto, April 5-7,, 1987? Sure would be convenient to have all the listed references available at the call of a mouse-click, on a CD-ROM disc.
Conclusion
The Art of Human-Computer Interface design is a new testament to the fact that Apple Computer, Inc. maintains, to some degree, the original fresh outlook, spunk, and populist vision that the company was founded on. The book sets forth, in chapter and verse, the beliefs, hopes, fears, and speculations of those who will help design the next generation of personal computing tools.
It's a credit to the book's editor, Brenda Laurel, that the dozens of different voices in this book speak all in a similar tone of voice. While the book benefits from the input of diverse perspectives, the diverse perspective are harmonious enough to create common themes and threads that are revisited from different angles.
One of the common threads that runs throughout the essays in this book is that every effort must be made to remove any impediments that stand between the user and the computer. To maximize productivity and minimize frustration, we need to think about how to remove those specific impediments that stand in the way.
Applying this thinking to the book itself, a serious impediment to the widespread reading of this book is its hefty price tag. It would be no exaggeration to say that the ideas in this book are valuable enough to be distributed free of charge.
Stewart Brand, the moving force behind the Whole Earth Catalogs, wrote an inspiring piece in the Whole Earth Review two years ago on how ideas themselves want to be free. It's a very Western sort of attitude that says, "I own this idea. This notion belongs to me." No mind stands alone. We are collectively one mind, not many minds.
Civilization moves forward in direct proportion to the willingness of people to share ideas with no financial, emotional, or other strings attached. Such is the pure state of being of a computer user group, where advice, information, and expert opinions are offered free of the encumbrance of money. The more we can export this sharing ethic to other sectors of our lives, the better off we all will be.
Phil Shapiro
(The author takes a keen interest in human-computer interaction issues. He can be reached at: pshapiro@his.com)