Ben Shneiderman talks about the surprising benefits of designing for diversity
Ben Shneiderman is a professor in the Department of Computer Science, founding director (1983-2000) of the Human-Computer Interaction Laboratory, and member of the Institute for Advanced Computer Studies and the Institute for Systems Research, all at the University of Maryland at College Park. He is the founding chair of the ACM Conference on Universal Usability, which will be held November 16-17.
UBIQUITY: How would you characterize the amount of progress that has been made in usability for computing and communications in the last 10 years?
BEN SHNEIDERMAN: There has been progress in usability, but not enough. User expectations and dependency on technology have grown much faster than the quality of the user interfaces. I think that has more users being more frustrated, more of the time. We can do much better in making the technology more useful and usable to a larger number of people.
UBIQUITY: There's been talk about usability and user friendliness for a long time. Yet people say they don't see it. Do you sympathize with that kind of disbelief?
SHNEIDERMAN: Yes, I'm sympathetic with the disbelief and frustration. I think the optimists among us believe that it can be better and we do see evidence that well-designed products are well received in the market and make life better for people. I recently got a message from my printer that said, "printer restarted due to fatal system error at address 0c1ad118." I didn't know what to do, of course. So I rebooted, waited 12 minutes and then tried again. One survey of 6,000 users suggests that 5.1 hours per week are wasted by people struggling with their computer. So there's more time wasted in front of computers than on the highways and that makes it a national priority to do better.
UBIQUITY: Is usability more a science or an art?
SHNEIDERMAN: I think it has elements of both. We are developing a scientific foundation under the title of human-computer interaction that enables researchers to understand some of the principles of cognition and to make better predictions about user performance (http://www.acm.org/sigchi). There is also the practitioner's art that's often called usability engineering, which is developing techniques and tools that allow for much more effective systems development, both more rapid development as well as higher quality. Using software such as Visual Basic and other development tool kits enables developers to produce systems more rapidly. However, the ambition of developers and expectations of users have grown faster and we need a better effort at quality control. Other technologies have succeeded in offering reliable mass media with universal usability -- such as television, telephone and postal services -- where complex technological systems are made accessible to a large number of people. Sure, there are problems with the mail service and, sure, the TV sometimes goes bad, but more or less, most people can succeed in getting what they want. I don't think we've reached that stage in the design of software for general use.
UBIQUITY: Please tell me what motivated your upcoming ACM Conference on Universal Usability (http://www.acm.org/sigchi/cuu/) .
SHNEIDERMAN: I was trying to define the goal of universal usability and define the research agenda (Communications of the ACM, May 2000, pages 84-91). As I probed deeper into the topic, I came to see how much it affects many branches of computing: networking, databases, graphics, human computer interaction and even operating systems design. It gives us a wonderful new set of challenges for computer science research and development. As the conference program chairs (John Thomas and Jean Scholtz) developed the call for participation they were seeking to draw greater attention to these challenges so that more companies and more university researchers will participate.
UBIQUITY: Please tell us about your background.
SHNEIDERMAN: My background is in computer science, doing database, file design and optimization techniques. I moved towards human-computer interaction with studies of how people use computers to improve the design. I'm proud of my 1983 efforts to develop the interface for the hyperlink -- the idea of highlighted words in a paragraph which you click on to jump somewhere. It was a small contribution, but it's very satisfying to see how widely it has spread.
UBIQUITY: ACM had a role in that, right?
SHNEIDERMAN: Yes, because the July 1988 issue of the Communications of the ACM was about hypertext. We made a hypertext electronic version of that entire issue using Hyperties, which was commercialized by Cognetics Corporation. ACM sold 4,000 copies of the PC-based version of the Communications of the ACM. And it's that document that Tim Berners-Lee cites in his Spring '89 manifesto for the Web. We had called the highlighted phrases, embedded menus, and he was clever enough to call them hot spots or hotlinks. The concept of the link goes back, of course, to Vannevar Bush, who inspired other important contributors like Andy van Dam and Ted Nelson. But the particular implementation and the authoring tools that enable users to make these links conveniently was our work. The April 1986 paper with grad student Larry Koved appeared in Communications of the ACM (http://www.cs.umd.edu/hcil/hyperties).
UBIQUITY: Tell us about your work for the Library of Congress.
SHNEIDERMAN: Another important step was our collaborative effort with the Library of Congress to redesign the online catalogue from a system that required a one- to three-hour training course, to a simplified touch-screen interface. That experience reinforced my belief that design improvements can make dramatic differences in the utility and accessibility of products. Skeptics suggest that satisfying diversity requires you to dumb down the interface or to go for the lowest common denominator, but we believe they're wrong. Satisfying the needs of diverse users often leads to more creative solutions that are beneficial to all users. More recently we contributed to the design of the Web site for the American Memory project.
UBIQUITY: What are the backgrounds of the conference attendees?
SHNEIDERMAN: We're seeing computer scientists, designers, and commercial developers, plus some policy-oriented people. Mostly we are attracting innovative developers of new interfaces. One particularly inspiring example is Gary Perlman's paper. He developed a new version of the search interface for FirstSearch (www.OCLC.org) . His project was to develop the Web interface in English, French and Spanish and make it work on multiple browsers. The software architecture that enabled him to make that an easy development also gave him the mechanisms by which he could make versions for the handicapped, for children, for printing and, also, make modification easy when changes were necessary. I love that example because it shows how the stimulus of trying to design for multiple, international languages and browsers leads to an architecture that facilitates many good things. It's like the curb-cut scenario for the disabled. Do you know the story?
UBIQUITY: No, what is it?
SHNEIDERMAN: Curb-cuts in sidewalks were designed to enable wheelchair users to get around in urban environments. But it turns out that there are many other beneficiaries such as tourists with roller bags, delivery people with pushcarts, roller bladers, and parents with baby carriages. Another example is Alexander Graham Bell, who was developing the telephone for the needs of the blind. Once you start opening your mind to the needs of diverse users, you come up with innovative technologies that benefit many.
UBIQUITY: You're suggesting an interesting premise. You're saying that designing for diversity stimulates thinking about usability?
SHNEIDERMAN: Absolutely. Here's another example that I find really stimulating for researchers in computer graphics. If you have a 500K JPG image on a Web site, it may work fine for people on high-speed connections, but it's just too large for people on a dial-up line. So shouldn't the user be able to request a 50K or a 20K version of that image? Shouldn't we have the technologies that would appropriately crop the image or resize it to make it smaller, or use fewer colors, or just use key features? There are many, many interesting problems that emerge from the stimulus of satisfying users that want a smaller byte count in the Web page they download. What if users could say: I don't care what you do, but I want this next Web page in five seconds. That changes our design of networking and of operating systems to have hard, real-time requirements that force reductions in the volume of data that needs to be transferred.
UBIQUITY: On the theme of shortening things, let's go back to the hyperlink. Could we imagine a URL descriptor that would be substantially shorter? In other words, instead of an impossibly long string, is there some way to give it a shorter name?
SHNEIDERMAN: Sure. There are some services that provide shorter, more memorable names. America Online, which deserves a lot of credit in trying to support universal usability, simply uses what they call a keyword instead of the URL. People learn those keywords and that makes for less typing and a more satisfying experience.
UBIQUITY: Suppose one wanted to cite the "Financial Times" which has an enormously long URL.
SHNEIDERMAN: You've identified a very common problem. Bookmarks are one solution to that, right? But the problem is that bookmarks quickly become very numerous and people have a hard time navigating them. Those are the kind of little usability problems that need greater attention. I'm not going to solve every one of them in this interview. Greater attention needs to be given to exactly these small annoyances that trouble each of the 100 million Web users every day.
UBIQUITY: How do you get the attention of those who should be solving these problems?
SHNEIDERMAN: I've proposed that every time that your machine crashes, you should get a dollar from the supplier, and every time you get a dialogue box you don't understand, you should get a nickel. It could be a credit towards future purchases. We need to find clever ways to make things like that happen. I went to a restaurant and the waiter spilled some soup on my pants leg and not only did they offer to pay my dry-cleaning bill, but they also gave a free dessert to all four of us at the dinner. We need to have a way that the suppliers of software and hardware are more accountable and we also need better forms of reporting. Just as we expect airlines to let us know the number of delayed and canceled flights, I think we should have a better understanding of the level of frustration and struggle that users have. This will encourage companies to develop better software and to compete on reducing the struggles and frustrations that too many users have.
UBIQUITY: Are there any major breakthroughs that you could hope to find?
SHNEIDERMAN: I think the main, unexplored notion is level-structured design. I would like to have interfaces begin with one-tenth of the complexity of the full system and the user could have a large slider, along the right-hand side of the screen, to move from level zero to one to two to three to four. As they move the slider up they would get more of the menu items and more help screens. The zero level would have a nice tutorial and no error messages -- simply allowing the basic functionality in a game-like manner. Game designers have successfully found ways to make users progress from novice to expert, gracefully, and sometimes, literally, in the levels of the game. I think having a level-structured design would be a real opportunity in many cases.
UBIQUITY: If you were giving out universal usability awards to commercial products, who would get one?
SHNEIDERMAN: Well, Quicken from Intuit, and America Online are a good start. I also like Adobe Photo Deluxe, which took a very complex graphics environment and cleverly enabled users to do many tasks in a simplified way. Similarly, PowerPoint, while not simple, supports an amazing range of powerful functions. Another favorite is the Macintosh, which was based on a good set of design principles. And on the Web: Yahoo and Amazon.
UBIQUITY: OK, let's end with an invitation for people to attend the conference.
SHNEIDERMAN: Good idea. This conference (www.acm.org/sigchi/cuu/) is ACM's response to the "digital divide," in that we're looking at strategies for reducing it. Even if technology were free, we'd still have to design it better to make it accessible to a wider range of people. We hope to stimulate researchers, developers, product managers and marketing people to understand that if they seek to satisfy the needs of novices and experts, young and old, and speakers of different languages, poor readers, disabled users, and elderly users, that they have an opportunity to create highly-successful products that provide real service to larger numbers of people. We think that e-commerce, government services, healthcare, and education online will all become greatly facilitated if more people think about the principles of universal usability.