Dissecting the hyperchunk

By David Walker (Google profile)

Time for a trip to the Web medium's deep past, and the early days of this column. Back in 1996, I reviewed a little volume called "The Future of the Book", a 300-page collection of interesting and learned conference papers that obsessed about print's fate in the face of the CD-ROM and the Web. The conferring theorists fretted that screens would replace books, that hypertexts would replace linear narratives, and that images would replace words.

All of which proves how tough is forecasting. In 2001, no-one bothers with any of these concerns. CD-ROMs contain software or games, but not books, because we all know that reading long documents on a screen is hard work, and screen technology doesn't threaten to change that anytime soon. "Hypertexts" - texts that offer multiple alternative paths through a narrative - simply don't exist outside the relatively narrow world of a few computer games. And one of the few trends in the mostly static world of Web design is to take the images off the screen and just leave the "real information", the triumphant word.

In place of the hypertext, we have what might best be termed the "hyperchunk" - the succinct slice of high-quality information designed to be stored in a database, linked from all over, understood quickly and in relative isolation, and often acted upon within a few minutes.

Why must a hyper-chunk be high-quality? Because it must survive a Darwinian struggle for attention and ranking among Web users and Web search engines. A hyperchunk must address its subject powerfully and honestly, even if that means addressing a very small niche topic. The alternative is to be overwhelmed by higher-status, more popular documents just a click away. (My own output as a columnist output is assessed by the Fairfax newspaper group partly on the basis of Web site traffic.) The typical writer or publisher also has to contend not just with competing writers but with practitioners who have suddenly discovered the Web as a fine cheap publishing tool. If you're writing about economics, for instance, you have to face the fact that fine modern-day economists like J. Bradford DeLong and Paul Krugman have Web sites that are only a click away, and that most of your readers will find them eventually.

Why must a hyperchunk be succinct, taking 250 words where a newspaper might take 1000, or 800 where a magazine might take 3000? Because on the Web, readers must work harder at their tasks; the sheer difficulty of screen-reading is now well documented.And because on the Web, people have the constant choice of beginning another experience through a hyperlink, bookmark or typed URL.

Why must a hyperchunk be self-contained, able to be understood in isolation? Because its readers can come from anywhere - from the article which preceded it, or from an outside link created years later in a publication whose readers will struggle to understand its meaning. And because the hyperchunk is typically stored in a database, pre-summarised and pre-indexed through "metadata", ready to be pulled out and used over and over again as circumstances demand. The point of the presentation must be obvious to viewers. It must also be obvious to the content management software that places it on a Web pages, and to the search engine that analyses that page.

Why must a hyperchunk often be capable of leading to action? Because the evidence of actual use shows the Web is a utility medium, whose users are often trying to solve a specific problem or perform a specific task. Good hyperchunks often owe a debt to direct marketing copy.

Some writers object to the idea of shrinking articles, of catering to information systems, of writing text designed to provoke action - and fair enough, since print is not likely to die soon. Many more writers have reason to fear a medium that pits them against true experts. But for others, hyperchunks offer the challenge of creating the Web's first literary sub-genre. Whoops - was that a forecast?