Issue
Article
Vol.29 No.3, July 1997
Article
Issue
Issue |
Article |
Vol.29 No.3, July 1997 |
Article |
Issue |
I have a number of great things to say about the Web '97 conference. First, it was in San Francisco, and the weather was warm and sunny. Second, the conference was well planned with excellent speakers and interesting courses. Third, and perhaps most importantly, Web '97 included two new conference tracks, usability and architecture design and maintenance. This is a major milestone for two closely related and traditionally under-appreciated fields.
In the usability track, I attended two very different courses, the first by Alan Cooper and the second by Jakob Nielsen. While there's no substitute for being there, I'll attempt a brief summary of each presentation.
Alan started out by proclaiming that he's a "software designer", not a "user interface designer" and that interface design is a subset of software design. With that clear, he proceeded to explain that the design of Web sites involves exactly the same problems as the design of interfaces for desktop computing.
Alan's presentation focused on two key problem areas. First, today's Web browsers are extremely limiting from an interaction design perspective. He chastised Web designers for accepting the tiny "sandbox" to play in that companies like Netscape and Microsoft have provided. Because of this problem, he explained that "most contemporary Web design is of little long term value".
Second, Alan went into detail on the topic of navigation design. Based upon the assumption that "people don't like choices", he explained that you must "inflect the choices" by creating a pyramid and that the simplest, most effective way to improve navigation is to provide fewer places to go. He added that you must help users to maintain situational awareness as they move through the site, employing the metaphor of fighter pilots who are told "lose sight, lose the fight".
Alan noted that hierarchies organize information around a single attribute but that people should be able to browse for information via multiple ad hoc associations. Things became a little fuzzy here as Alan began to talk about what he calls "monocline grouping". The best substance I could pull from this portion of the discussion was that our interfaces should support a piles and files metaphor to allow for a more visual and chronological mode of information retrieval.
Jakob introduced the notion that successful design requires elements of both art and science. He correctly observed that most Web site design involves pure art and leaves out the science.
Jakob proceeded to explain the type of science he practices, namely user testing or usability engineering. He used the design and redesign of SunWeb, Sun's internal Web site, as a case study to illuminate the process and techniques behind rapid, highly iterative, relatively low cost user testing. Since most ideas from the presentation are covered in detail on the Sun Web site (see User Interface Design for the Web), I won't repeat them here.
I was intrigued by Jakob's discussion of the redesign of Sun's internal Web site. Apparently, the old design was falling apart as department after department ignored style, architecture, and navigation guidelines, opting to reinvent the wheel which each sub-site. According to Jakob's calculations, the cost of inconsistency at the page level was costing Sun roughly $1 million per year and at the site-wide navigation level $10 million per year.
These calculations were based upon estimates of pages and sites visited per day per person × a few seconds of confusion each due to differing navigation options × the numbers of staff at Sun. Jakob admitted these were just ballpark figures. Hey, we've all got to walk that fine line between art and science sometimes.
The redesign involved development of a standard navigation bar with 3 rows as follows:
Row 1:
SunWeb logo, Sub-site name and logo (required)
Row 2:
Navigation buttons: SunWeb Home, Search (required)
Navigation buttons: What's New, Site Map, Next/Last Page (optional)
Row 3:
Hierarchy Bar showing the full path the user has taken from SunWeb
Home to the current location (required)
The navigation bar Jakob presented was obviously well thought out and well designed. However, I left wondering whether the creative, rebellious people throughout Sun would really follow even the best designed style guides. Only time will tell.
In summary, Alan Cooper certainly presented some thought provoking (albeit not very new) ideas about information retrieval and interface design. However, he didn't back up most of his ideas with examples, and I left feeling rather confused as to exactly what he was proposing.
Jakob Nielsen, on the other hand, backed up all of his ideas with evidence obtained through the testing of real, live users. As one of the few speakers who has experience in the science of Web site design, Jakob deserves respect for introducing usability testing into the wonderful world of the Web.
Finally, this conference provided an opportunity to consider the relationship between information architecture (as seen from the perspective of librarianship) and more traditional interface design. I disagree with Alan Cooper's argument that the Web introduces no challenges that haven't already been tackled on the desktop.
On my desktop computer, I label and organize relatively small volumes of information for myself. On the Web, we must label, organize, and index large volumes of heterogeneous information for many users. These challenges invite the types of insight and experience found in the fields of information science and librarianship.
I was very pleased to see both usability and information architecture recognized and represented at the Web '97 conference. I hope that we are entering a period of more enlightened Web design that involves interdisciplinary teams of experts who each bring their own unique strengths to the table.
Jakob Nielsen
http://www.useit.com/jakob/
User Interface Design for the Web
http://www.sun.com/sun-on-net/uidesign/
Alan Cooper
http://www.cooper.com/staff/staff.html#top
Peter Morville
http://argus-inc.com/staff/
morville.html
Argus Associates
http://argus-inc.com/
Web Architect
http://argus-inc.com/design/webarch.html
Peter Morville is vice president of Argus Associates, a firm that specializes in strategy consulting and information architecture design for complex Web sites. He holds a graduate degree from the University of Michigan's School of Information and Library Studies. He has participated in the design and production of Web sites and Intranets for Barron's magazine, Borders Books & Music, Consumers Energy, Dow Chemical, SIGGRAPH, the University of Michigan, and many more organizations. Recent publications include the Internet Searcher's Handbook and a monthly column called "Web Architect" in Web Review magazine.
Peter Morville, Vice President, Argus Associates, Inc., 109 Catherine Street, Ann Arbor, Michigan 48104 USA,
tel: +1.313.913.0010,
fax: +1.313.213.8082,
morville@argus-inc.com
Issue |
Article |
Vol.29 No.3, July 1997 |
Article |
Issue |