|
Too many links or too little thinking?
January 3, 2002 Editors fail readers by putting too many links on newspaper site home pages, asserts a recent article published by the august American Press Institute. I disagree. Complaining about too many links on a home page is like complaining that a piece of music has too many notes. In the December issue of the American Press Institute's new online magazine, Steve Yelvington, vice president of Web development for Morris Digital Works, counts links on newspaper home pages, finding that the Miami Herald has 244, the Chicago Tribune 243, and the NY Times 211. He asks: "Were there really 200 or so top-level choices that I desperately needed to make?" Yelvington pauses to condemn "lack of organization," but who doesn't? Most of his energy goes into counting and complaining about links. Counting links, like counting notes, misses the point. Sure the Miami Herald is a visual cacophony, but is this because of its 244 links? After all, Yahoo.com does an eloquent job distilling the web's several billion pages into a comprehensible grid of categories. And Yahoo, which has spent millions of dollars on usability testing, has 245 links on its home page. What really destroys a home page is not lots of links, but links that are poorly organized, categorically muddled, imbedded in a kaleidoscope of colors like... yes, it has to be said, www.americanpressinstitute.org, home to Yelvington's article. (Here's a challenge: go to that home page and try quickly to navigate to Yelvington's article without using the search function.) Although the API site offers 80% fewer links than the sites Yelvington critiques, its 50-odd links are a muddle, arrayed in 10 or 12 multi-colored clumps, each with a hazy conceptual and visual border. Visiting the site is like trying to shop in a Walmart that's been stocked by a monkey: soup cans are shelved with toy guns, tire chains with baby oil. In fact, too few links often hide lazy thinking. Sites that offer only a few links impell the visitor to play three card monte, clicking back and forth and up and down through the site trying to guess where the webeditor has hidden information. Of course, careful selection of category headings can help. But it is even better to show readers some or all of the subcategories hidden beneath these these top level labels; remember the Journalism 101 dictum to "show not tell"? The result of more information is enlightenment. Like Yahoo, NYTimes.com's 211 links are easy to scan and digest because they have been broken into intuitively obvious (or at least well-defined) groups and subgroups. To offer scientific buttress for the view that too many links spoil the stew, Yelvington invokes a famous paper called "The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information" which demonstrates, Yelvington says, that "the eye and brain can't easily process clusters of more than seven choices." Since this paper, published by psychologist George A. Miller in 1956, is one of the totems often invoked by designers to sanctify their obtuse choices, it is worth reading closely. (See this article for another deconstruction of Miller's misuse by interface designers.) Yes, Miller concluded that people cannot, in a 1/5 second flash, accurately count more than seven objects. Moreover, seven seems to be the most pieces of stimuli humans can process at one time in a "unidimensional" environment. But Yelvington must have missed the Miller's next point: "You may have noticed that I have been careful to say that this magical number seven applies to one-dimensional judgments." People can process far more stimuli, Miller wrote. Viewing marks displayed in a two-dimensional grid, suddenly "people can identify accurately any one of 24 positions..." Bump the "dimensions" up to six (by adding color, pitch, depth, etc), and Miller estimates that people can distinguish among 150 different stimuli. Given that site editors can guide readers with words, fonts, font sizes, headers, colors, columns, and decks, and that these readers can draw on past web-reading experience, a homepage's total link count is irrelevant to its usefulness. What matters is how well editors group and display these links. As Miller concludes, "By organizing the stimulus input simultaneously into several dimensions and successively into a sequence or chunks, we manage to break (or at least stretch) this informational bottleneck." Remember Yahoo's 245 links? They are broken up into seven major groups. Perhaps that is today's lesson. |
Featured sites
Services
|














