New Ways of Organizing Data Could Change Nature of Computers

Published: April 4th, 1997

Scientists' designs are based on how people tend to remember and classify information

By Jeffrey R. Young

The computer "desktop" started out as a tidy metaphor for keeping track of a few documents in a handful of folders. But as people do more with their computers, and as the machines hold more information, many desktops have become disaster areas -- crowded with overlapping windows and with folders on top of folders inside of other folders. Information ends up lost in the electronic shuffle, and users forget which application is doing what in which window.

It may be time for a better desktop, or perhaps for a whole new metaphor for organizing information on personal computers -- maybe one that's a little more user-friendly. Consider these:

At the heart of these projects is the idea that the current framework for organizing electronic data doesn't take advantage of the mind's ability to make connections among disparate pieces of information. "We're underutilizing our remarkable human possibilities," says Ben Schneiderman, director of the University of Maryland's Human-Computer Interaction Laboratory.

Another goal of the projects is to let users spot patterns and make new connections within a vast amount of information. Researchers hope that putting more on the screen at once will allow whole constellations of data to emerge, helping chart the way in the universe of information.

They agree on one thing: The desktop as we know it is overdue for improvement. The new systems take advantage of the way the mind conceptualizes things rather than the way paper has always led us to organize information. "We've sort of been locked in this menus-and-windows metaphor," says James D. Hollan, head of the computer-science department at New Mexico. "I think we're seeing a movement toward more-active information space, where we're not mimicking the printed page."

Another researcher puts it more bluntly: "The desktop is dead."

The idea behind Lifestreams is that the things people do on their computers form a kind of digital diary. "Lives are time-ordered, so your electronic life should be also," says David Gelernter, a computer-science professor at Yale who invented the system.

The software does away with divisions among different kinds of documents. Every file -- whether electronic mail, a word-processing document, or a spreadsheet -- is part of the same personal-information stream. There's no need to make up a name for each new file and decide where it fits in with others. New documents are simply identified by date and time.

In a prototype, documents are displayed as pages stacked like an infinite hand of playing cards, with the top left portion of each page exposed. Selecting a document brings it into view for reading or editing. Users can also rely on a powerful search tool to sort through the stream. Documents identified in searches, or substreams, also appear in chronological order.

"It really helps people deal with information overload," says Eric T. Freeman, who recently completed his doctorate at Yale by working on the project. "People remember when something happened."

Dr. Freeman has put most of the 22,000 documents he's created in the past few years on the Lifestreams system. This has helped him quickly locate such information as a newspaper article about Lifestreams requested by a reporter. "If I had to find it in a traditional file structure, it would take me all day," he says.

Now he wants to bring a commercial version of Lifestreams to a software store near you. He's helped set up Lifestreams Inc., and expects the company to offer a retail version of the software within two years. In the meantime, researchers have set up a World-Wide Web site that offers technical information as well as screen shots of the software in action (here).

To some people, of course, a lengthy time line seems far more cumbersome than the familiar desktop. Dr. Gelernter acknowledges that the Lifestreams system might not work for every user. "I've never claimed it's right for everyone," he says, "but it's right for me, and it seems pretty clear that there are a fair number of people with the same general requirements I have."

Pad++ (pronounced "pad plus plus") doesn't exactly replace the desktop metaphor. It still relies on the idea that an electronic storage area can be thought of as a kind of space. But the system takes the spatial analogy farther than the desktop ever has, extending the workspace beyond the confines of the computer monitor. Here, your workspace becomes an endless surface, and the monitor acts as a viewfinder. It's as if you were peering down on an electronic world below.

Users get around this space by what is called a "pan-and-zoom" interface. By backing up or zooming out, they can look down on a whole field of documents, the pages looking like small squares and the words as illegible blurs. Zooming in slightly makes major headings readable; zooming closer still brings a particular document into clear view, so that it can be read or edited. You arrange documents on this surface the same way you might lay out on a table the notes and other materials for a paper you're writing, grouping them conceptually.

Jonathan Meyer, a research scientist at New York University who is working on Pad++, says the system taps into people's remarkable ability to remember things based on location. "People have vast spatial memories," he notes. "You remember how to get home at night, and you do it without much trouble at all."

With a system like Pad++, people could use visual cues to remember where on their computers they've put their papers or notes, he predicts. "In Pad, you can fly around and say, 'I remember that rectangle.'"

Of course, providing a larger space can make it easier for some users to get lost in cyberspace. In early versions of Pad++, users could zoom in so close that they seemed to be looking through a magnifying glass -- not the best way to find information spread out over a large area.

"That's one of the things we're working on," says Mr. Meyer, who works under the direction of Ken Perlin, the N.Y.U. professor who invented Pad++.

Benjamin Bederson, an assistant professor of computer science at the University of New Mexico who is also working on the project, says Pad++ is particularly useful for navigating the World-Wide Web. As users move from Web page to Web page, the software can help create a visual map of where they've been, instead of the simple list of visited sites that current browsers present. Unlike a list, such a map might reveal relationships among the sites you've visited.

Researchers have set up an on-line slide show to demonstrate Pad++ (here).

At Maryland, Dr. Schneiderman oversees a number of computer-navigation projects, among them Elastic Windows. In most current systems, only one window is active at a time. In this system, however, several windows are visible and active at once, joined on the screen like floor tiles. As one window is enlarged, other open windows shrink. The program lets you move data from one application to another far more easily than is possible now.

Having so many open windows can be confusing at first, and Dr. Schneiderman acknowledges that "it does take users a few minutes to get acquainted with the idea."

Another research project at Maryland is called "TreeViz." It can scan a hard drive to present a single, graphical picture of its contents. "It gives you X-ray vision to see your whole hard drive at once," says Dr. Schneiderman. A quarter of the files on most hard drives are duplicates or documents that are no longer needed, he estimates. The software is designed to help people root out this deadwood and free up more storage space.

A list of projects at the University of Maryland's Human-Computer Interaction Laboratory is available on the Web (here).

Researchers say the tools they are working on could have a considerable impact on how people work with computers. The systems could give expert users more -- and easier -- control over their electronic lives. And Dr. Schneiderman and others hope that the tools can help simplify computer use, so that newcomers will have an easier time becoming comfortable with the machines.

As Yale's Dr. Gelernter puts it, the goal is "to make life less of a nuisance. Make the complexities and depredations of the modern world less of a pain. Give us more control over our lives, and more time to go fishing."


Copyright 1997, The Chronicle of Higher Education. Posted with permission on the Pad++ web site (http://www.cs.unm.edu/pad++/press/chronicle.html). This article may not be published, reposted, or redistributed without permission from The Chronicle.