Our ongoing love affair with paper

Paper - messy, bulky, high-maintenance (it must be collated, filed, retrieved and cared for in between), and, in sufficient quantities, a fire hazard. Technologists have long wanted to get rid of it. How much cleaner and neater a world it would be if everything was stored electronically to be searched for and retrieved at the touch of a couple of buttons.

But in the real world, every office that gets a computer in the belief that it will reduce the use of paper discovers the contrary. This was true in the days of batch-card processing, when vast printouts were needed to debug programs and inspect results. It was true in 1981, when the first personal computers were launched, making document copies easy to replace. And it is true now: the consumption of office paper more than doubled worldwide between 1980 and 1997.

Abigail Sellen and Richard Harper, who in the late 1990s were researchers at Xerox's Cambridge EuroParc lab, decided to look into this anomaly. What causes our apparently undying affection for paper, despite its many inconveniences and costs? The usual answer from technologists is that screen technology is not good enough. The contrast ratios of paper are better, so it is less tiring to read. Consumers are likely to point out that paper is portable, does not need batteries and can be dropped in the bathtub without electrocuting the reader.

Sellen and Harper's exhaustive research observing people in their offices shows, however, that these are the least of the reasons why paper endures. Most people, when reading material for work, make notes, add comments in the margins, highlight material they want to return to and read across several documents at once. All of these things are easier with paper. While today's word processors support annotations, these are often hard to pick out from the surrounding text or require effort to access. Computers are poor at supporting the kind of "layering" that happens when several people in turn read a document and add notes in different colours.

Similarly, we forget how many clues we derive from a paper document as a physical object. You can tell at a glance how far into the document you are; you can refer back and forth between two pages by keeping a finger on each one and flipping the partial stack. This kind of thing is doable with computers, but reading Sellen and Harper's observations reminds you how frustratingly clunky it all is. Who has not felt lost in the middle of a 100-page electronic document, found using bookmarks annoying, or found it distracting to have to switch among windows to read and annotate?

Yet much of this material has never been researched before because technologists have tended to frame the question in simpler terms: how to make better screens, for example. Sellen and Harper instead observed paper use within organisations, including the UK police force and the International Monetary Fund, and ultimately a selection of 15 people with differing jobs, to examine how and when they read and how and when they used paper. The reading diaries the subjects kept revealed that 85 per cent of document-related activity time was spent on paper-based reading and writing. Observations of organisations, on the other hand, revealed that often they did not understand how their own employees used paper. In a chocolate manufacturing company, for example, it turned out that the departmental, supposedly shared, files the company wanted turned into a central repository, were in fact idiosyncratic collections of material generally accessed by, and understandable to, only one individual. The real knowledge was not in the paper and so could not easily be transferred into digital form: it was in the people's heads.

However, this volume is not just a hymn of praise to paper. Sellen and Harper devote the final quarter of the book to working out how to incorporate a greater understanding of how people read and use paper in order to design better technology to support those activities. As they show early on, reducing paper use is successful only if it is part of redesigning work processes. Paper is not the problem, it is a symptom. In their police study, for example, forces found their new database unusable because the way it handled form-filling did not support typical interactions with the public - even though reverting to filling out forms on paper doubled their workload. The digital world could support many of the goals that paper supports now, and in such a way as to create the best possible combination of the two.

Wendy M. Grossman is author of net.wars and From Anarchy to Power : The Net Comes of Age .