| WebWord.com > Interviews > Stranglehold of the Priesthood (2-Aug-99) |
Conducted via email by John S. Rhodes (2-Aug-99) What is AskTog? AskTog is, first of all, my free webzine, where I get to blow off steam about design issues that are important to me and, I believe, to the industry. AskTog is also the title of each month's lead column, where I try to come up with answers to issues raised by my readers. While the webzine is centered on human-computer interaction, I do wander somewhat afield, covering everything from human-golf ball interaction ("The Gentle Learning Curves of Miniature Golf") to aviation human factors ("When Interfaces Kill: What Really Happened to John Denver"). And, when you tire of user interfaces altogether, you can read about such misadventures as my diving, fully dressed, into a freezing cold pond in Holland ("The Would-Be Hero").
Writer, publisher, chief cook, and bottle washer. Pretty much everything except for editor, which duties are sporadically performed by the estimable John Scribblemonger. As I explained to Mr. Scribblemonger only last Thursday, upon spotting yet another split infinitive, I would perform that duty to were it not for my day job as lead designer at a little startup called Healtheon/WebMD. In the "old days," I would write a book every four or five years, summing up what I'd learned at my day job. AskTog is my next "book," evolving on the web now, rather than springing into print several years down the road. The web is a much more exciting, immediate medium.
That's a bit like asking why do mammals stay up all night. Only nocturnal animals and teenagers stay up all night, and only graphic designers and industrial designers are likely to care more about how an interface looks than how it works. Graphic designers typically receive zero training in usability. Industrial designers do receive at least some exposure to usability, but many learn to rise above it. A case in point are those worthies that decided it would be cool to make the iMac mouse perfectly round, ensuring that users would have only one chance in 360 of holding the mouse at the correct angle. Interaction designers, on the other hand, care exclusively about how things work. They spend so much time caring, in fact, that they tend to drive everyone around them crazy with their complaints about everything from the design of refrigerator door handles to iMac mice. Most interaction designers, unless they are trained in graphic design, absolutely depend on their graphically-talented counterparts. Likewise, most graphic designers and industrial designers have enough sense to depend on the talents of interaction designers and ergonomists. A few just must strike out on their own, usually in both senses of the word.
Interface design is usually very, very expensive. That's because interface design is usually done shortly after the product ships. If, however, you are clever enough to do it at the beginning of the project, it can save you time, effort, and lots and lots of money. A good designer can develop a competent, implementable design in short order. It's not magic; it is a proven methodology: Come up with a preliminary design, test it on potential users, refine and retest until done. After sufficient iterations, if you have have carefully chosen your user population, and if you have been honest with yourself about the outcome of the tests, you will achieve a successful design. The savings brought about by such a procedure begin to pile up quickly. How many times have you sat in a meeting with 10 engineers, each burning $100 or more per hour in salary + overhead, guessing at what users want? Those futile meetings cost companies tens of thousands of dollars with no return on investment. With a competent design in place, such meetings disappear. First-customer-ship of an untested product will invariably start the phones ringing. Lots of money disappears into the tech support maw. Then the panic sets in, and the engineers are quickly back having their guessing sessions. Version 1.1 rolls out, followed by 1.2 and 1.3, each inching ever so slowly toward the same design that could have been put in place before a half-million lines of code were written.
This is a better question for the CEOs of AOL and Compuserve. AOL figured having a usable interface would be a good thing. Compuserve did not. Compuserve has folded it's unfriendly little tent and became an ISP, while AOL has enough money in petty cash to buy Compuserve and a large portion of South America.
The users are exactly the same. What is different is that traditional GUIs are based on nineteen-eighties personal-computer technology, while browsers are mired in nineteen-sixties time-share technology. Browsers, lacking such advanced features as menu bars and usable dialog boxes, have presented web designers with almost insurmountable hurdles in trying to implement anything but the simplest functionality. In 1997, I spent five months just trying to get new users successfully through the process of centering and enlarging a new window, since the 3.0 browsers insisted on opening new windows half way off the screen (Maximizing Windows). When you are expending more than half your resources trying to get around the bugs in the "operating system"—in this case, browser—it eats up precious time needed for actually designing your own weblication (web application).
Good design has enabled more people to come to the party. The automobile never took off until the electric starter permitted other than strong men to drive a car. It never became universal until the automatic transmission freed people from being a peripheral to an infernal combustion engine. In the same way, the graphical user interface freed people to interact with their task, instead of their machine. As comfortably seductive as was the intricate keyboard maneuvering of the glass teletypes of CPM, ProDOS, and MSDOS, it was grossly inefficient for many of the domains that the visual interface has opened up, such as graphics. As web browsers become mature, they will enable people to similarly interact with the world without dealing with the artifacts of the medium. Computers and the Internet have been around for a long time. The big change is that the stranglehold of the priesthood has been broken. Interaction design has played a central role in crushing the monopoly, to the benefit of everyone, including the former priests. As browsers work their way out of the primordial ooze, they will cause a secondary revolution, as today's primitive forms-entry weblications are replaced by weblications that cover tasks previously the exclusive province of "real" applications.
The most "critical system operation" in a human being involves coordinating a smooth rhythm of the heart muscle. The conscious control required to ensure critical system operations of a computer should require the exact same level of conscious control. Computers are ridiculously complicated. Using a computer is like commuting to work on an industrial tractor, festooned with a hundred different unlabeled levers and switches. Humans should perform no system functions. Ever. Humans should make choices. I want to go to Fresno. I want soup for dinner. Humans should not be required to reconfigure their engine so their car can drive toward Fresno or screw in a different frequency magnetron in their microwave oven because this particular soup is incompatible with the magnetron just used to melt the butter. The web is setting us free. No longer do we have to send in our $24.95 for a diskette with this month's upgrade to FileMaker, then wait six weeks for it to arrive. Soon, we won't even have to click 4000 "Next" buttons to get though the installation. We will make the decision that we want our software to remain updated, and it will happen automatically. (This is already happening today in isolated circumstances, but it is far from universal.) This level of automation comes with a terrible responsibility. You can't shut off all reasonable access to the "basement" of the computer unless your automation really works. Few things are as intolerable as mad wizards that can't do the job and won't get out of the way so you can take over.
Usability problems arise from two sources: People failing to test their software and people testing it, but then failing to take corrective measures. Failure to test tends to occur out if ignorance—a remarkable number of developers still don't understand the importance of usability testing. The second failure tends to occur out of ego—"I know my software is right; the test subjects just must be stupid."—or out of expediency—"We'll fix it in the next release." The underlying systems themselves also offer significant usability limitations. In the case of GUIs, what were well-designed interfaces back in the early 1980s have been overwhelmed by the sheer power and complexity of the task domains of today. What works for a two-page memo writer doesn't necessarily scale to cover the needs of a 200 page multimedia document or six million page FDA drug application. The biggest problem with the web—and the problems are myriad—is its invisible navigation. Users typically don't know where they are, where they came from, or where they are going. Individual designers do their best to offer menus, breadcrumb trails, maps, and a variety of other idiosyncratic cues and clues. However, this information should properly be supplied by the system, so that all web pages offer consistent, recognizable navigational cues.
The era of good interface that existed in the eighties ended ten years ago and will remain dead for at least for the next ten years. The computer industry is following the lead of the aviation industry, paring back to the absolute minimum in comfort and usability consistent with not losing customers. Unlike the airlines, the computer companies have not shrunken the size of their seats nor plasticized their food. Instead, they have just stood still, investing little or nothing in new interface technology while the world rushed ahead all around them. Just as with the airlines, this course has paid big dividends, with minimalist companies like Microsoft burying companies like Apple, with its former emphasis on quality. (The difference between Microsoft and Compuserve is that Microsoft had sense enough to cobble together a bad GUI interface while Compuserve just sat on their hands and waited for the end to come.) The web, meanwhile, will continue to improve upon its retro beginnings, eventually gaining some parity with traditional applications. The next big breakthrough will involve a fundamental shift in input and output. 300 dot per inch and above screens will gradually replace today's course monitors. A successor will be found for the mouse that enables more than one or two bits of information to flow from user to computer (click left, click right—some vocabulary!) Likely some melding of gesture and voice control will rid our screens of much of today's visual clutter. With high resolution and enriched interaction comes the possibility of 3D, with users swooping over information spaces, zeroing in on "locations" with the same ease and precision people can accomplish today in the real world. The second revolution will be born. Thanks Bruce. Great stuff.
|
If you want to know when
new interviews go online, |
Home | Moving WebWord | Cool
Books | Hot Web Sites | Reports
Newsletter Archive | Services | Interviews | News | About WebWord.com
URL: http://www.WebWord.com/interviews/tog.html
© 1999 by John S. Rhodes. All rights
reserved.
Do not reproduce or redistribute any material from this document,
in whole or in part, without explicit written permission from John S. Rhodes.