useit.com -> Recommended Websites -> Spotlights -> 2000 Q1-Q2 Spotlights

Spotlighted Links From January-June 2000

Please note that some of these links may have suffered linkrot. They did work as of the date when they were recommended, but the Web is a highly fluid medium.

June 24, 2000
Scott McCloud (author of Understanding Comics and Reinventing Comics) has started a monthly column on visual communication on the Internet. Despite being (ahem) all graphics, this column is worth the download time and proof-by-example that it is possible to use visuals to overcome users' reluctance to read online. No blocks-of-text syndrome here.
June 23, 2000
They learn young: At the recent Human Factors and the Web conference, Terry Sullivan from the University of North Texas presented a study of children using the Web. Half were 12 years old; half were 16. 94% of the kids exhibited banner blindness and ignored graphic promotions, even when they contained a direct link to the thing they were trying to find. When asked why, one kid said "oh, I saw that graphic, but it looked like an ad, so I ignored it." Same answer we have gotten so many times when testing adult users. Some differences: This latter finding conflicts with the belief that kids will just keep clicking and are less intimidated by technology than adults. The Sullivan study is only a first stab at studying Web usability for children (only two age groups, only 16 test subjects, only two websites), so we need more data before drawing firm conclusions, but I do find it striking that these findings contradict much of the folklore around children and the Web.
June 20, 2000
British Telecom claims a patent on hypertext links based on Prestel in the 1970s. Contrary to most such claims, there is some basis for BTs patent. The history of hypertext shows that BT did not invent hypertext: the first running software was the The Hypertext Editing System and FRESS at Brown University in 1967, and the conceptual invention was even earlier (1945 or 1965, depending on whether you credit Vannevar Bush or Ted Nelson). But BT may have a case for hypertext links across a network (i.e., the Web) since all the early systems were mainframe based. I can't tell without spending several days decoding the legalese language in the claims.
June 17, 2000
Gnutella does not meet any established usability criteria (exception: very easy to install).

The user interface is totally mysterious and does not guide the user towards achieving his or her goal. Instead, it exposes too much technical detail (IP numbers and the like) which will be totally confusing for the average user. The design is bragging about its own technical prowess as opposed to helping users do things easily (the term "user illusion" comes to mind - would be an appropriate concept to improve Gnutella).

Just one example: when you start Gnutella for the first time, it doesn't do anything. You are faced with a screen with countless options that are not explained very well, and clicking on the buttons does nothing. Only if you know the magic incantation will anything happen. You have to type the address of a server in a certain field which is fine if you:

Not mainstream software.

This is a classic example of software designed by programmers without the involvement of any human factors experts or technical writers (no help included in Gnutella). It has obviously been very successful (1.7 million copies downloaded), and there is a large group of users who are willing to suffer through the user interface because they want to download free music. If you are a college student with too much time on your hands and five nerds in your dorm to help, then Gnutella works just fine.

The entire open software movement is run by programmers who are motivated to bring out advanced code and not motivated to simplify the user interface to make it approachable by less-technically inclined mainstream users. If they want hundreds of millions of users (as opposed to a few million), it will be necessary to fix the user interface and bring it up to the standards of usability expected of professional software.

(Given this critique, I should also say that Gnutella is revolutionary on a different level than the miserable surface UI. The deeper analysis is that these new applications reconceptualize the structure of the Internet from point-to-point connections to a true network and involve the individual users more closely in constructing the available services. Of course, a conceptual breakthrough will count for nothing if it is presented in such a difficult manner that very few people can use it.)

June 15, 2000
Great article about the Japanese i-mode system for mobile Internet access. I am particularly pleased that DoCoMo (the company behind i-mode) is starting to make multi-billion-dollar investments in Europe and the U.S. Let's hope that the Japanese can save the rest of the world from WAP.
June 12, 2000
Staples has redesigned their site to make it easier to use. By making a small change to the way they asked users for address information, Staples reduced the number of users who would simply leave the site by 75%. Great example of the value of following usability advice. And also an example that relatively small elements of the user interface can have big impacts on the business performance of a site.
June 12, 2000
According to Fortune, only 9 of the 25 sites with the most traffic in Europe are actually European. Admittedly, they use "unique visitors" which is a bad way of measuring traffic and user loyalty, but the table still indicates the current American dominance of international Web use.
June 11, 2000
I have come across one of the exceedingly rare examples of good content for the Web: Computer Sweden runs a 60-70 word micro-column on their home page every day (with a link to a bigger article for more detail). See the area called "Kommentaren" with a B&W photo of the columnist. Most of you will not be able to read the text, but the authors are usually quite good at encapsulating one good perspective on a major story or issue into a nice Web-optimized content nugget.
June 6, 2000
Worst home page of the year: MeT Briefing. Obviously, the sponsors (Ericsson, Motorola, and Nokia) are so poor that they could not afford to get a usability professional to look at the site for about one minute - all it takes to generate a list of glaring problems: And what's with the funny animal anyway?
June 3, 2000
Salon has now fixed the two design sins I listed here on May 26: they have stopped forcing a fixed font size on readers and they now have links from the home page to the featured articles. Kudos for the quick fixes of these flaws. There is nothing wrong in making mistakes (even though these specific mistakes ought to have been caught by usability before launch); the only thing that's wrong is not fixing the mistakes. In fact, I have personally made almost every mistake I warn people about in my books and articles.
June 2, 2000
Frequently asked question: We are a web design shop. How can we sell our clients on user testing? Won't they think that we don't know our jobs if we have to test?
June 1, 2000
A British developer's experience with WAP (semi-technical).
June 1, 2000
A Flash site with an editorial explaining why most use of Flash is bad. To return the favor, let me add that some use of Flash is good; it's just rare to come across appropriate use of Flash on the Web today. I was amazed by the animated logo on this site. It came up fast and stopped moving before it became annoying. Why can't everybody follow such simple rules?
May 31, 2000
Jesse Berst correctly notes the weakness of the AOL/Gateway Web Pad: it comes with a phone cord so that you can't carry it around the house. Portability is one of the main benefits of tablet computers. Most press coverage has focused on the use of a non-Windows OS and non-Intel CPU, but who cares about the guts of the machine. The entire idea of Internet appliances is that the users won't even know that there is such a thing as an OS humming away somewhere inside the tablet. What matters is the UI, and the biggest question regarding the AOL/Gateway appliance is whether AOL "owns" the user or whether the device encourages freedom. (Disclosure: Nielsen Norman Group has lots of stock options in a Web tablet that is cordless.)
May 31, 2000
Good discussion of WAP on Slashdot. Note in particular comments from a Finnish programmer (nobody actually uses WAP in Finland; they just talk about it) and the multiple comments describing why the Japanese i-mode system is better than WAP (10M users in Japan; unfortunately the rest of us may never get it).
May 27, 2000
Comments on Salon's redesign: most people think it's bad. Two mistakes they could fix quickly: an HTML sin (fixed-height fonts that don't resize to fit the user's preferences) and a hypertext sin (no direct link from the home page summary of a story to the story itself - users are forced to navigate through an intermediate page).
May 26, 2000
The Web is slowly moving towards pay-for-content. Latest service is Themestream which pays authors 10 cents for each page view their articles attract. As an experiment, I have put my essay "Ten Biggest Thinkers of the 20th Century" up on Themestream. So if you click on the link, I get 10 cents. The missing element is obviously a way for Themestream to make their money back: there is no way that an undifferentiated page view is worth more than a fraction of a cent on an advertising basis. So the current business model seems to be to funnel money from gullible VCs into authors' pockets. Good.
I don't believe Themestream is viable since a collection of random essays is not very compelling. Even if users are attracted by some good essays (like mine?) they will not convert into loyal customers who turn to Themestream first whenever they want to read something. Internet users don't want to read random commentary, they want to solve problems and get specific information. Targeted services are the way to present content on the Web.
Even though Themestream is doomed, the service shows the way forward by providing a business model for authors: once you get paid for writing, people will develop better-quality Web-specific content. In the future, micropayments will provide a way to bring pay-for-content to all sites and sites that solve real problems for a targeted group of users will win.
May 18, 2000
Many people ask me about the best places to work in the user interface field. One company to look at is Trilogy Software of Austin, Texas. I know the people in Trilogy's human-computer interaction group, and they are good (including one of my former colleagues). They are looking to fill HCI positions in usability, interaction design and visual design. They also are hiring HCI managers. The company develops e-commerce software for big clients (more info in a recent Interactive Week article).
May 17, 2000
Boo.com has closed. Good riddance. Boo was one of the very few high-profile sites to launch in recent months that dared violate my design principles and aim for glitz rather than usability. Contrary to some analysts, I don't think that Boo should drag other e-commerce sites down with it. The collapse of Boo does not prove that e-commerce doesn't work. It proves that overly fancy design doesn't work. I do agree that the story ought to teach investors some lessons about reading the business plan. They need to check two things: If these two issues are not spelled out in the business plan (or if the answers are the wrong ones), you will most likely be wasting your investment. Users just don't want to take it anymore.
See also my review of Boo (written in December 1999) and a former Boo employee's lessons learned.
May 12, 2000
Debunking a product for computerized assessment of websites: "A Web Site User Model Should at Least Model Something About Users," by Peter Pirolli, a senior Xerox PARC scientist who invented the concepts of information foraging and the "scent" of a site. I applaud Dr. Pirolli for taking the time to use his superior knowledge and brainpower to set the record straight. Unfortunately, trade magazines have mainly relied on the companies' own press releases in describing several voodoo usability services that have been released in recent months. Before websites waste money on such services, they are well advised to check the insights of the leaders in the user interface field.
May 10, 2000
A new study from the Pew Internet & American Life Project concludes that the Internet does not make users lonely after all. After criticizing the Stanford study that said that the Internet makes us lonely, fairness dictates that I point out that the new study has its own weaknesses. Most fundamentally, the study relies on a telephone survey: that is, calling people up to ask them questions instead of observing their actual behavior. Thus, all we have is self-reported claims and not true data. For example, 59% of respondents who exchange email with a family member claimed that they are in contact with that relative more often thanks to email. True or false? We don't know. Even if we trust people to remember everything and to report accurately, it is not always clear what one can conclude from survey findings. Another example: 72% of Internet users say they visited family or friends the day before the survey, while only 61% of nonusers report they had visited someone. One explanation for this finding is that the Internet makes people less lonely. Three other explanations are Despite these weaknesses, I tend to believe the main conclusions of the new report. But we are still waiting for the ultimate study that looks at what people do and not what they say they do.
May 10, 2000
SOAP is being considered by the Web Consortium: supports coordination between software running on different platforms using XML transferred across the Internet. A step towards richer user interfaces than asynchronous browser-server interactions. In the spirit of the Web, SOAP works cross-platform (whether "platform" means operating system, programming and scripting languages, or application software).
May 5, 2000
Two good usability conferences have announced their programs:
May 5, 2000
After trying for three hours to find somebody to give my $3,000 I conclude that it is impossible to be allowed to buy a laptop on the Web:
May 3, 2000
How not to do it: Thinking about buying an IBM ThinkPad 570? Want to look at the specifications? Even though this is the second-most important link (after "buy it"), it is in tiny font and low-contrast color. Assuming you find the link, you are taken to a list of all IBM laptop specifications, from which you have to remember the product number of the machine you wanted. Find 570 on the list and click on it: what do you get? Not the specs, but another page with a link to the specs. Clicking this link finally does download the specs, but adds insult to injury by making it a PDF file. Considering that all I really wanted to know was three numbers (physical dimensions of the machine), I only read 15 bytes out of the 149,821 they force me to wait for. Three clicks instead of one and ten thousand times as slow a download as needed. Sorry, IBM doesn't get the $3,000 I have budgeted for a new laptop. Advice to anybody who does want my money: put your product specs on a regular Web page and not in PDF - not only is HTML faster to download, but the resulting text is easier to read on the screen (PDF does look better when printed and is the perfect format for distributing an entire product brochure, but that's not what people want when they need to check one or two numbers in the spec sheet).
April 25, 2000
The Government's plan for breaking up Microsoft has serious problems. It does make sense to split off Office into a separate company since Office is the true monopoly. Kicking it out would have two benefits: But the divestiture plan also includes a suggestion to split Internet Explorer from Windows and put it into a company with MSN and the Internet content division. Clueless: If splitting up Microsoft, keep Internet Explorer in the Windows company.
April 21, 2000
I have seen the future and it is (gulp) Interactive Barney. Probably the most dangerous product shipped by Microsoft in recent years: partly because it is so good and partly because it is a persuasive technology aimed at changing little kids' minds. Interactive Barney is one of the case studies in the new book Information Appliances and Beyond, edited by Eric Bergman. This chapter explains why computerized plush animals introduce a whole new range of usability problems. The book has many examples showing why the next generation of user interfaces will finally stop cloning the Macintosh. This book is both visionary and practical: future consumer electronics, toys, and games need great usability or customers will abandon them. "My new cell-phone is so complicated that I need a two-week training class." Not!
April 20, 2000
Very impressive: Adobe published their interview with me in five languages simultaneously: English, French, German, Japanese, and Swedish. This level of commitment to international content is still rare, but will become the rule in the future for major corporate websites.
April 17, 2000
Photos from the recent CHI conference in The Netherlands (warning: slow download).
April 14, 2000
I have updated the portal referral statistics to reflect traffic in March 2000. Biggest news is strong growth of Google and AskJeeves. Google is now the third-biggest source of referrals. Yahoo remains the largest, but "only" increased by 46% last year (compared with a 114% growth in overall referrals). These stats are a likely leading indicator of the future prospects of navigation services.
April 14, 2000
Internet World UK reviews the blue/green book in the March 2000 issue and concludes: "If the Web design company you are employing hasn't got a copy of this book on its shelves, you'll know to go somewhere else."
Good point :-)
In fact this is good advice. Never hire somebody who doesn't know the rules, even if you are after the wildest bleeding-edge design. There are indeed cases where it is appropriate to break the rules, but only if you know them.
April 1, 2000 (but not a joke)
ICONOCAST reports on two estimates of the amount of time spent on the Internet every month by users in the United States: This is a great example of why one should not rely on surveys or other user-reported "data." People cannot remember how they spent their time (at least not in detail: the users' recollections were 75% off the true value, which may not be that bad for an estimate). Why did people think they spent more time on the Internet than they actually did? Maybe it's because it is considered "good" to be online, so people tend to over-report the extent to which they use the Internet. Or maybe the story is just that it is hard to estimate time (which has been documented in countless studies).
March 21, 2000
Bell Labs has done it - developed more bandwidth than even I think we need for the perfect Internet user experience: three terabits per second. In 1995, I computed that we only we need a single terabit. Of course, a family with four users would saturate the new Bell Labs line if they were all going full steam simultaneously.
March 14, 2000
Steve Ballmer (Microsoft CEO) gave a great talk at the PC Forum conference today where he emphasized the need to give users more control over the Internet and recapture the balance of power in favor of the user's tools rather than the websites. I commented that even though I agree with his vision, I found their products sorely lacking in the area of user empowerment. For example, IE 5 does not have better user support than Mosaic did 7 years ago. Ballmer answered that the next few releases of software would probably have to continue along the original weak lines, but that more empowerment-oriented software should ship in about 3 years. Well, I have always said that IE 8 would be the first good browser :-) The San Jose Mercury News has more details about Ballmer's talk.
March 11, 2000
Register to vote for the future of the Internet. ICANN (the main governing body for the Internet) will have an election for their board of directors later in the year. Any Internet user can vote, but because they want to verify each voter's physical existence, the password needed to vote will be mailed out by physical mail. Thus, it is necessary to register far in advance if you want to participate in this important election: the first real democracy in the governance of the Internet. Voting rights are available to all individual members of ICANN, and it is free to join (and they have very strong privacy guarantees). It took me about a minute to register: Go to members.icann.org and register now and you can help make the Internet a better place by voting for candidates who care about usability and users. See also: New York Times article about the election (access to the Times requires free registration).
February 25, 2000
ICONOCAST (great email newsletter) computed the cost-per-visitor from running commercials during the Super Bowl (biggest annual TV event in the U.S.):
1998: $3
2000: $38
This is the cost to get somebody to take one look at your home page. With a conversion rate of 2% (generous), the cost per customer would be almost $2,000. A clear indication that it is a failed strategy to prioritize name recognition above user experience.
February 25, 2000
Bruce "Tog" Tognazzini lists some of the advances in user interfaces that have already been invented but are still not found in incremental designs like Apple's OS X and Microsoft's Windows 2000. Time to go beyond the UI research from the 1960s, he thinks (I concur). Note how the Web could also be much richer if we could find ways of more tightly integrating improved PCs with smarter servers. The IBM mainframe user interface for VM running on 3270 terminals was amazingly good, but not really the best model for e-commerce, universal information systems, and virtual company workflow support.
February 24, 2000
John C. Dvorak has an interesting design critique of a leaked draft of MSN's next generation. It is cluttered and suffers from an unwillingness to link and allow users to move around in an unconstrained manner. Dvorak claims that an inside source told him that the new MSN design was driven by "politics by pixel" - the higher up on the ladder you are, the more pixels you get on the home page for your pet project. If this is true, then it is a prescription for usability disaster. Too bad: AOL needs competition and it would seem to be easy to build something better now that AOL's design is driven by who pays how much and not by user needs.
February 14, 2000
Stewart Alsop correctly criticizes Sony for copy protection in its new Music Clip. Copy protection failed for software, and it will fail for content. You want to have fewer barriers between your customers and your services; not more. The Music Clip is even more of a paradox since it is intended to play MP3 clips: first, users download normal-format music files from the Web; then they have to spend extra time to convert the clips to a proprietary, copy-protected format.
January 30, 2000
Epinions announces that it has passed the 200,000 reviews mark. Very good for a new site in less than five months. About a week ago, Google announced that its traffic grew by 70,000% in 1999. Both results show that it pays off to focus on users when launching new Web services.
January 25, 2000
For about a week, Word magazine redesigned its home page to look like Yahoo (it's now back to looking busy). Even though it is good to see more sites embrace simplicity-based design, the experimental home page for Word contained multiple usability problems. It is not enough to copy Yahoo's appearance. Ironically, doing so perpetuates the misunderstanding that Web design is mainly about how things look. Yahoo's design does not work because it's two columns of text, some of it bold-faced. It works because it has an appropriate integration of content and interface. The Yahoo-like Word home page contained several problems:
January 19, 2000
Bruce Tognazzini reviews the Design of Apple's new "Aqua" user interface for OS X. Tog says that it makes for a great demo but violates many usability principles. Worst of all, Apple wasted the opportunity to design a substantially improved GUI. Instead, they are "dipping the 20 years old Mac/Lisa UI into the 10 years old NextStep UI and covering the result with a thin layer of candy." A good way to treat apples at county fairs, but not a good way to improve the lot of users who suffer under information overload and need drastically increased usability and productivity.
January 17, 2000
Invention disclosure: Fitts' Law Sensor for auto-hiding docking areas. We all know that the taskbar in Windows'98 exhibits extremely annoying behavior in auto-hide mode when you want to point to an object toward the bottom of the screen. You just can't point to things without activating the taskbar. New behavior: only activate the unhide feature if the user is banging the cursor against the edge of the screen with a velocity above a certain user preference. If you are trying to select something at the edge of the screen, you will be moving the mouse slowly and carefully and will not activate the edge of the screen, even if you happen to point to a pixel at the edge.
January 5, 2000
Usability review of Canada Post's new ePost service.
January 4, 2000
Study of European users found that trust was enhanced when the design was simple. The author speculates that people feel more secure when they feel more in control of the experience.
January 3, 2000
An art critic says that 3D films are less suited for conventional storytelling and more suited for "the drama of pure form in pure space", comparing 3D IMAX to architecture, dance, and absolute music (in a NY Times article that is no longer online). I have said for years that 3D is ill suited for most user interfaces (which decidedly need plot and agency) and better suited for object display.
January 1, 2000
Happy New Year! We are on our way to a Year 2100 problem :-) The real issue is a usability concern: the desire of humans to save time and space and only use two digits for the year - as shown by the main newspaper headline for the first day of Y2K.
NY Times headline: 1/1/00

Other Months in 2000

Other Years