New Perspective

This review was originally published in the June 1996 issue of Windows Tech Journal. and is reprinted here with permission.

Two recently published books provide innovative, thought-provoking approaches to thinking about and designing the face your program presents to the world.

ABOUT FACE: THE ESSENTIALS OF USER INTERFACE DESIGN

Alan Cooper is the original designer of Visual Basic, a product that has had a tremendous impact on our notions of how to create software. In About Face, Cooper takes the elements that make up a graphical user interface and examines them individually and collectively. Even the most familiar gadgets get a fresh look.

About Face comprises 34 chapters grouped into eight sections. The book begins with broad design issues and then homes in on specific idioms. The index is superb; almost every anecdote and topic is listed. I enjoyed the book’s loose, organic structure. The writing is friendly, conversational, and easy to read. Unfortunately, many points are repeated excessively. A good edit could have trimmed much of the 550 pages without losing an idea, insight, or example. The text is also sprinkled with unusual words such as spoor, epony­mous, and tessellation—the latter two on the same page. Such unfamiliar words tripped me more than once and broke the flow of communication.

Still, these flaws are minor. All in all, the book is an enjoyable read.

TIPTOE THROUGH THE TOOLTIPS

Cooper’s views often contradict conventional wisdom. He offers convincing evidence and examples, however, that illustrate the poor design of much of today’s software. He argues, for example, that “transliterated mechanical models are always worse on computers.” One example he cites: calendars. Paper calendars usually hold one month per page because of the mechanical limitations of paper. Computers suffer no such limitation, and, Cooper asserts, it makes more sense to provide a contiguous calendar in which each month merges smoothly into the next. “The most insidious problem with metaphors, the real show-stopper,” he writes, “is that they tie our interfaces to mechanical-age artifacts.” When was the last time you heard someone sound such a note of caution about GUI metaphors?

It’s interesting that Cooper often uses analogies to everyday objects such as cars, boats, and airplanes to support his main points. The design of jet cockpits, for instance, demonstrates the importance of keeping controls in or around a user’s field of vision. For computer software, this field of vision corresponds to the main window and surrounding area. In Microsoft Word, for example, buttons for common actions are located directly above the main window; various progress indicators are in the status area below the window. This is one of those ideas I had been peripherally aware of but hadn’t consciously considered. About Face offers many similar observations.

Cooper has a knack for seeing subtle characteristics that make something either useful or cumbersome. He considers tooltips to be “one of the cleverest and most-effective user-interface idioms.” Although balloon help on the Macintosh is similar, he notes, it contains full-sentence descriptions of everything on the desktop—scrollbars, menus, status bars, and so on. This arrangement may be useful for first-timers, but it’s intrusive and condescending for expert users.

What we need, Cooper argues, is a taxonomy: a way of classifying the elements and characteristics of user interfaces. Then he proceeds to provide one throughout the book, sometimes invent­ing words to describe a category. Butt­ con, a contraction of button and icon, seems ill considered. What’s wrong with calling them toolbar buttons?

Cooper’s taxonomy is more valuable when he divides programs into four classifications: sovereign, transient, daemonic, and parasitic. Sovereign programs are used for long, continuous stretches; word processors and compilers are examples. Transient programs perform a single function and are then closed or minimized. Daemonic programs carry out background tasks, and parasitic programs combine aspects of sovereign and transient programs. Parasitic programs are superimposed on other programs—RoboHelp, which is a Microsoft Word add-in, is an example of a parasitic program. The benefit of this classification is that the author offers guidelines for design decisions based on a program’s classification. A transient program, for example, should not take over the entire display.

PUT ON YOUR DANCING SHOES

About Face uses a couple of interesting special notations. Axioms are boxed and displayed in large type. In contrast to the main text, axioms tend to be bombastic and absolute. Some typical statements are “Good user interfaces are invisible,” “Never make the user ask to ask,” and “Never bend your interface to fit a metaphor.” These broad assertions serve to define the boundaries of the discussion, and they certainly encourage thought.

The other notational device is design tips. Like the axioms, the design tips are scattered throughout the book. These pithy rules of thumb include “Give modeless dialog boxes consistent terminating commands,” and “The program should perform optimally on hardware that doesn’t exist yet.” The design tips are less inflammatory and provide more useful and practical advice than the axioms offer.

Cooper’s book shows him to be clear-minded, innovative, and visionary. You’ll find many original, thought-provoking, and illuminating ideas in About Face. Whether or not you agree with everything Cooper says, you will come away from the book with a deeper understanding of interface issues and new insights into the needs of software users.

GUIDELINES FOR ENTERPRISE-WIDE GUI DESIGN

Written by Susan Weinschenk and Sarah Yeo, Guidelines for Enterprise-Wide GUI Design offers advice for developing applications in Windows, OS/2, and Motif environments. The main audience for this book is companies seeking to establish enterprise-wide standards for their programs’ look and behavior. This intended audience isn’t surprising, given the authors’ backgrounds. The bio on the jacket identifies Weinschenk as a consultant specializing in GUI design. Yeo is the manager of technical documentation for Compuware Corp., a consulting firm.

Standards are intended to enforce consistent appearance and behavior among all applications in a department, site, or enterprise. The use of standards eases the tasks of software developers. Because many decisions have already been made, developers need not worry about as many style considerations. Standards also promote consistent implementation of common situations. If a standard isn’t provided, each developer is likely to lay out a form or position a control differently. Standards also make a user’s job easier. Users can apply what they’ve already learned from the company’s other applications—and from commercial applications—when they en­ counter new programs.

Thumbing through GUI Design, I experienced a joy from my grade-school days—the book is loaded with pictures, illustrations, and screen shots. What could have been a dry set of arbitrary rules is instead clearly and compellingly illustrated. When an idea differs from one operating system to another, examples are provided for OS/2, Motif, and Windows 3.x. The book has surprisingly little text—only enough to explain the examples. At 222 pages, the book is tight and sticks closely to the subject.

GUI Design is organized into 14 chapters and three appendixes. The book covers a range of topics, including controls, menus, metaphors, navigation, and interaction. A so-called Customization Guide is intended to help readers adapt the authors’ guidelines to the requirements of their organizations. The chapters present a blend of graphic-design techniques, writing-style guidelines, and usability principles. The authors also discuss current GUI conventions, tips for using the various elements of a GUI, and ways of translating everyday objects into computer application metaphors. The wide range of topics does not allow in-depth coverage of any topic, so I was happy to find Appendix B, which provides a reading list. I would have found that information even more useful had it been footnoted in the text. Still, the book provides quick answers to a number of interface questions.

IMELDA HAD NO AIR JORDANS

The advice given in the book ranges from obvious to arbitrary to obscure. In the category of the obvious, the authors recommend that you use a “non-destructive” default button and, for menus, create unique mnemonics. Among the arbitrary assertions is that interfaces should have no more than three colors per screen. The obscure statement “Show three to eight items at a time” appears to reflect the notion that humans find it easiest to consider seven options, plus or minus two, at a time. Although most of the guidelines make sense, it would be nice if the book offered more justification for some of the guidelines.

The book’s longest chapter is devoted to online help. Writing text for online viewing differs significantly from composing a manuscript to be read on paper. Because the screen is the medium, Weinschenk and Yeo note, you should try to keep topics brief and self-contained. It’s best if users do not have to scroll through a topic. Among other subjects, this chapter covers types of windows, the button bar, keyword searches, and glossaries. Con­ceptual and organizational considerations are also examined. The authors write that users generally turn to online help to answer immediate problems; online help should contain action information rather than general information. If you are new to this area and are confronted with designing a highly interactive, cross-referenced help system, this chapter alone may be worth the price of the book.

THESE BOOTS WERE MADE FOR CUSTOMIZIN’

The authors’ goal is to provide you with a significant bundle of standards that you can customize to reflect what is unique to your organization. The book offers suggestions for putting together a company standards-development team and presents ideas on rolling out, implementing, and maintaining your company standards. Many of the suggestions are obvious. The authors state, for instance, that “for guidelines to be successful people must buy into them. The best way for people to buy into guidelines is to be involved in the process.”

GUI Design is intended for companies that want to establish their own GUI design standards, but any programmer moving into GUI design would benefit from the information assembled here. The book is eminently practical.

If you want to know more about the theory of interface design—if you want to know the why as well as the how—then Alan Cooper’s book will interest you.

JUST DO IT

The software world is filled with user interfaces only their designers could love. A little thought and care, combined with the knowledge and ideas in these two books, can go a long way toward making users’ lives easier and more productive. If you apply these principles to your interfaces, your program will fit your users like those long-ago sneakers fit your feet by the Fourth of July.

Austin Avrashow is president of Alpha Documen­tation, a technical documentation and Windows Help development company based in Portland, Oregon.