New Way To Work

In 1987, Apple Computer produced a short promotional video, imagining how people would use technology at the dawn of the 21st century. Its anticipation of the way professionals would one day work was remarkably prescient. In the video, a globe-trotting professor uses a portable computer to download data from a worldwide information network. An intelligent software agent becomes his liaison to online information. He speaks commands to his machine in plain English and authors an interactive, multimedia document for colleagues to share.

Not everything outlined in that scenario has come to pass, of course. The tablet computer heralded by the video evolved into the Apple Newton, a spectacular failure, and speech-recognition software still isn't widely used. But the Internet, software agents, and multimedia collaboration have steadily infiltrated our daily work habits, and advancing software and falling computational costs are chipping away at barriers that have made talking to PCs and entering handwritten notes into them fantasies for most workers.

Peer inside the research labs of some of the world's largest technology and office-equipment companies, and new visions of the office of the future burn brightly. Among the heady ideas: sensors that detect when we enter our offices and cue software that alerts colleagues that we're available to talk; microelectromechanical systems capable of rendering super-sharp images on displays the size of a wall, working in tandem with tablet computers that network to the shared-team display; and business software that helps users find and assemble information by learning relationships among the words and phrases people use.

Which of these dreams of a future office will become a reality in the next decade? No one can say for certain. But because today's prototypes and predictions often become tomorrow's buying decisions, some IT managers are paying close attention to the ideas under discussion. "Four years ago, we didn't focus on emerging technology," says John Thomas, CIO at Parsons Corp., a Pasadena, Calif., engineering and construction company with 11,000 employees. As a result, Thomas says, Parsons dealt with too many small, unstable IT vendors whose products didn't stay current. Now, the company has four emerging-technology specialists who are among Parsons' highest-paid and most talented staffers. "We had to get more acquainted with our vendors' vision," Thomas says. "We like to influence that vision whenever we can."

Workspaces Get Smarter
Thirty miles from Manhattan, up the Saw Mill River Parkway, between a strip mall and a lumber yard in Hawthorne, N.Y., IBM and office-furniture maker Steelcase Corp. are building the office of tomorrow. Or, more precisely, one cubicle of it. But what a cubicle.

Overhead lights, which respond to chips in the ID tag IBM Research senior manager Marissa Viveros wears, change from blue to green as she enters the workspace, signaling her arrival. Other electronic sensors respond to her entrance by illuminating a task light above her desk, resetting the temperature to her comfort level, and triggering a message to her team members, via IBM-designed Web software, that she's back at her desk. A monitor on an outside wall of the 11-foot-by-7-foot workspace, which had indicated that Viveros was at a meeting, changes to show that she's now available. But if Viveros chose to do so, she could also temporarily "hide" from the system. "Customers want this dichotomy of privacy, combined with the ability to interact with their colleagues in open spaces," she says.

BlueSpace, as IBM and Steelcase call their creation, tests a few trends that will shape the workplace of the new century. One is increased collaboration. Another is employees' ability to handle more information-the idea that younger workers accustomed to managing multiple streams of information simultaneously will use those skills to more closely monitor their work environments, and in the process, work more efficiently as individuals and as teams. In the BlueSpace cubicle, for example, a touch-screen beside the user's primary display tracks team members' whereabouts and availability, and cues instant messages for colleagues.

"The nature of work is changing just as dramatically as the technology is changing," says Mark Greiner, who served as Steelcase's CIO for 3-1/2 years before taking over the company's tech ventures and research efforts last year. Steelcase's customers increasingly want to figure out how to make employees' workspaces mesh with their office technology.

Jeff Austin, VP of innovation in the corporate real-estate division of Wachovia Corp., a Charlotte, N.C., bank, concurs. "More companies are going to open workspaces, and that has an impact on technology and an impact on the way teams work," he says. "There's a lot more collaboration in the workplace, but also the need for concentration." As Wachovia consolidates its offices in the Southeast and reassigns workers, Austin is advising IBM and Steelcase, a supplier to the bank, on BlueSpace, and researching how its concepts can help his company make real-estate, IT, and human-resources decisions in tandem.

Austin's team also is investigating how Wachovia can measure each worker's overall cost and increase his efficiency. "BlueSpace is about trying to shift from cost-per-square-foot to cost-per-employee," he says. One conclusion: Employees who can avoid unwanted interruptions and cut time spent looking for colleagues are more productive and, so, less costly. "The technology cuts to the chase of all that," Austin says.

BlueSpace may help solve the problem of how to encourage communication and teamwork without stripping away personal space, an idea whose time may have come and gone along with the New Economy. "Teamwork doesn't have to mean just take all the walls and doors down and throw everyone into an open space," says Jack Whalen, a principal scientist at Palo Alto Research Center, recently spun off from Xerox Corp., who studies the impact of technology on work. "Companies tend to get enamored with this idea of 'We're all equal, and nobody has a desk.'"

More than 100 customers of IBM and Steelcase have toured the BlueSpace installation at the Hawthorne lab and a duplicate BlueSpace cubicle at Steelcase headquarters in Grand Rapids, Mich. The vendors say they plan to test BlueSpace with perhaps a half-dozen customers this year.

How much would a cubicle like BlueSpace cost an IT manager? The setup at IBM's Hawthorne Lab contains about $5,000 worth of equipment. "You can spend that on cherry furniture," says Michael Karasick, IBM's chief technical officer for pervasive computing. "It depends where you put your money."

Change can sneak up on us, and some ideas that have been on the table for a couple of years may soon see wider acceptance. In December, RNL Design, a Denver architectural firm that specializes in commercial and government projects, completed an eight-month exhibit in downtown Denver called Workplace 2010, built to illustrate how future office-scapes could give companies more freedom to reassign workers and recommission space.

One proposal: A raised floor to accommodate cabling that lets workers plug notebook computers and IP phones into access points around the office, to get to the Internet and networked printers. Combined with mobile furniture, the design fosters teamwork by letting users set up ad hoc workspaces when necessary, grouping and regrouping as projects and business priorities change. "The idea of increased flexibility and reduced cost of change is pretty critical to nearly every organization right now," says Amy Tabor, a senior associate at RNL Design.

As architects, IT vendors, and office suppliers try to offer companies integrated solutions-a la BlueSpace-issues arise about who makes the buying decisions and who absorbs the costs. "Those are the big questions," Wachovia's Austin says. "What I applauded about BlueSpace is the same thing I feared-it changes the way we do business." Austin's group at Wachovia is charged with addressing both workspace and related technology questions. "That's the model of the future," he says. "It's just getting there from here that's the question."

Collaboration Support
When knowledge workers sit at their desks, they have a wealth of data and resources at their fingertips: E-mailed memos, online calendars, Web sites, and reports. They also have the means to share that data with colleagues. But meeting rooms are a different story. Unless you use Bluetooth wireless technology to zap data from notebook to notebook-and almost no one does-sharing digital information at a typical meeting is a pipe dream.

The IBM-Steelcase BlueSpace project and others under development may put that dream closer to reality, ushering in a new wave of collaboration. "The most important work going on in any organization today is going on in some project room, some team room, some jury-rigged space where people can get together," says Mark Greiner, head of technology research at Steelcase. The BlueSpace cubicle will offer workers a collaborative environment. It includes a projector that beams E-mail messages and other documents to a wall or conference table for shared reading. The goal: To let employees mark up those documents using their fingertips, which could trigger motion sensors in the projector to draw a line, circle a key point, and so on.

Gary Starkweather, a software architect at Microsoft Research, has even more ambitious ideas. He envisions boosting collaborative endeavors by creating a large "people-sized" computer display that can be viewed and updated by all the participants in a meeting. Such a display could work like an office printer, using a software controller that parses input from multiple computers on a network so the output can be shared by a group. Starkweather knows whereof he speaks: In the 1970s, he invented the laser printer at Xerox's Palo Alto Research Center; that, coupled with colleague John Warnock's PostScript software, enabled WYSIWYG printing from myriad computers.

One challenge to revolutionizing display technology, Starkweather says, is building cost-effective readouts that don't sacrifice resolution as they grow in size. Enter nanotechnology called microelectromechanical systems (MEMS), which lets scientists fabricate microscopic mechanical devices on silicon chips. MEMS work by projecting light through an array of tiny lenses-about 2,500 on a 6-millimeter-square part. Behind the lenses sit a series of MEMS shutters, each representing a pixel. As each shutter is activated, it can turn light on, off, or partially on. A second array of lenses captures the light and creates the on-screen image. Starkweather says manufacturing billions of MEMS devices would eventually result in costs perhaps as low as $1,500 per display.

Microsoft and its hardware partners are preparing other contributions to enhance the way meetings are conducted. Due to ship later this year, Tablet PCs running a version of Windows XP that recognizes handwriting and speech input could transform note-taking and E-mail for users away from their desks, and one day they might even network to shared MEMS screens. In a recent E-mail interview, Microsoft chairman Bill Gates said tech-industry trends evident in Tablet PCs-longer battery life, low-power CPUs, and cheaper displays-"will have the same kind of impact that inexpensive memory and chips had. They'll work together to transform the way people think about computing."

Falling component prices could propel other advances in meeting technologies. Cheap storage underlies Microsoft's RingCam, an experimental device for recording meetings, built from off-the-shelf digital cameras and microphones worth about $300. A one-hour meeting recorded with the omnidirectional RingCam consumes about as much as a gigabyte of disk space, but with storage priced at about $1 per gigabyte, a company could store a week's worth of meetings for the price of lunch.

Office workers could benefit from RingCam by archiving reams of searchable data, says Microsoft senior VP Rick Rashid, who heads the company's research group. "You can get to the point where nothing you do as an individual has to go away, unless you want it to." While researchers are working out the kinks, software under development in the labs could let workers who miss a meeting browse the event for relevant information, using voice- and face-recognition algorithms to search for comments made by particular speakers.

But pulling video from servers hogs bandwidth, and IT shops need to lay the groundwork now for future applications, says Bob Egan, VP of IT at Boise Cascade, the $7.42 billion paper and lumber manufacturer in Boise, Idaho. "We want to get our network poised to use all the apps we think are going to become increasingly popular," he says. Last year, Boise ramped up a network upgrade project that aims to refresh the operating system and groupware on nearly every server in the company and prioritize traffic, to clear room for more online training and video conferencing. IT has to be proactive about network design, to avoid scrambling to put in place technology advances senior personnel hear about from other sources. Says Egan, "You're never as successful that way."

Protected Documents
Digital rights-management technology, which embeds in a file metadata about what's in it, who can see it, and how it should be used, got a lot of attention two years ago when Napster gave away MP3 files on the Internet and Stephen King published the first widely read E-book. But the law put an end to Napster's freewheeling music exchange, and King has gone back to paper.

But in the office of the future, digital rights-management technology could make documents traversing the ether more regulated, encoded with permissions about how they can be viewed, traded, updated, and printed, enhancing workers' ability to protect proprietary information and track the flow of their work. However, much of the technology to do this needs to grow up. Microsoft chairman Bill Gates acknowledged the problem at a Microsoft Research event last fall: "If you send out a document and you only want four or five people to see that, is it easy to do that?" Not really. It's also not easy to link information in a document to another data source so that it's dynamically updated. "It's the one category people somehow don't imagine we can do a lot better for the user," he admits.

John Thomas, CIO of Pasadena, Calif., engineering and construction firm Parsons Corp., wants to see Microsoft build some of that management functionality into its Exchange E-mail server. "It doesn't do a good job managing attachments and who can see them," he says. With 10,000 PC users exchanging contracts, blueprints, and other files, Parsons needs a way of associating documents with the E-mail they arrived with, controlling versions of files, and searching for them. But those functions don't exist today, he says.

Offices are "fertile environments for associating users' roles with permissions to read and distribute files," agrees John Erickson, a researcher at Hewlett-Packard Laboratories. Digital-rights-management systems piggyback on a company's security infrastructure to authenticate and authorize business users for reading, modifying, or printing files. That complicates matters, because colleagues who work for different companies are separated by firewalls. To address the problem, HP is working to help ratify rights-management standards and has joined Sun Microsystems' Liberty Alliance, which seeks to build cross-company trust systems.

Rights-management technology will likely scale better for enterprise apps than in the music and book business, where digital distribution is harder to swallow, says Michael Miron, CEO of ContentGuard Inc., a Microsoft-Xerox joint venture that develops and licenses software for digital rights management. "It's highly valued content, most of it not for sale, that needs the same protection as entertainment content," he says. ContentGuard licenses its Extensible Rights Markup Language, developed at the Palo Alto Research Center, to Microsoft, Xerox, and others. XrML and other digital rights-management software authorize access to content or a network service as it travels through cyberspace, in a language that multiple systems can read.

The technology has yet to find its way into tools for creating documents, such as Adobe Photoshop, Microsoft Word, and QuarkXPress. Doing so would pose some challenges for IT executives, Erickson says. In addition to problems implementing digital rights management across company lines, unchecked use of rights management could check the open exchange of information that characterizes working online. "If you have an unthrottled usage of digital rights management," he says, "it could potentially be inconvenient for end users."

Computer, Listen Up
The PC has gone a long way toward enhancing productivity-letting users quickly create and distribute documents, calculate the myriad costs of a product in development, and compile impressive project-management scenarios. Despite that, the PC still can't make the same mental connections that a longtime co-worker can to help us find information we need when we don't ask for it in quite the right way. "We can't reproduce the common sense of a 5-year-old child," says Michio Kaku, a professor of theoretical physics at the City University of New York. Nor can the PC fill in for an efficient assistant who can translate our scribbled notes or rushed comments into an electronic document.

Some relational capabilities are closer to reality on the PC than others. One of the more intriguing concepts is being studied by Microsoft Research's natural-language processing group. That team is building a giant brain-at least that's how they refer to MindNet, a database and set of algorithms that learn relationships between dictionary definitions and encyclopedia entries, in seven languages, acquiring more data as they work.

As researchers ask MindNet questions-How can I book a trip to Seattle? Who shot Abraham Lincoln?-the system learns about relationships between words and phrases, attempting to amass knowledge. If the researchers get it right, they could build new interfaces to the Web that help users get the right answers to search requests, even if they've posed a question too generally or too specifically, or chosen keywords a step or two removed from the spot-on choices. "On the Web, almost every fact you want is out there," says Eric Brill, a Microsoft researcher. "We just don't have a very nice interface to it yet."

Many savvy companies are inventing their own. Use of artificial-intelligence tools and expert databases jumped last year to 36% of companies in the InformationWeek 500-which lists companies with more than $1 billion in annual revenue judged to be the most innovative users of IT-compared with 26% in 1999. Twenty-seven percent of sites are deploying systems that profile employees' expertise.

The ability to make associations in a number of languages could open up more applications, such as searching for documents written in, say, Japanese, by speaking a query in English, and letting the system perform the translation work. Microsoft could also offer online services that let computer users publish factual text to the Web in XML format, where it could be grabbed and compiled by Office apps for writing reports. "Part of .Net Microsoft's Internet strategy could be that people publish information in a way that's very easy to access," Brill says.

How users will enter search queries, write E-mails, or annotate their documents may well change before they ever see commercial applications for MindNet. For all of the alternative graphical user interface metaphors that computer scientists have posited over the years-overlapping index cards, a tree, art galleries hung with documents, even Microsoft's Bob-the desktop-and-document system developed at Xerox's Palo Alto Research Center in the early '70s has proved remarkably resilient. It's evolving, though. Microsoft is building support for speech and handwriting recognition more deeply into the Windows operating system, and researchers at many companies are exploring how users can better communicate with Web sites and their PCs. Talking to the office computer could finally become more commonplace this decade, but there's a difference of opinion on what will be said.

Some experts argue that speech recognition will be used primarily in coming years for "command and control"-driving menus and opening files in conjunction with a mouse or other input device. But Chuck Thacker, a distinguished engineer at Microsoft Research and a principal designer of the pioneering Xerox Alto computer, predicts speech will be mostly used for dictation, because command and control applications offer supernumerary possibilities, from which a bad choice by the system can really louse things up. "I'm less optimistic about using speech for command and control," Thacker says. "There's more of a chance for something to go wrong or for something surprising to happen." That can confuse users-verboten in user-interface design.

John Hummel, CIO at Sutter Health, a $3.5 billion not-for-profit network of doctors and hospitals in Northern California, says the health-care industry would "go insane for speech recognition" if it were more pervasive and accurate. Computer voice-recognition capabilities are near the top of doctors' and nurses' IT wish lists, Hummel says-speaking surgery notes into a PDA instead of using a tape recorder could cut down on time-consuming and costly transcription services, for example. Radiologists want to create comprehensive electronic patient files that combine transcribed reports with digital images, which could speed insurance processing. But the 95% to 97% accuracy rate of current voice-recognition software needs to improve before it's "more trouble than it's worth," Hummel says. Sutter is conducting voice-recognition research and development with IBM, Microsoft, and other vendors, he says.

Speech interfaces and artificial intelligence could increase productivity, but, like many of the technologies that promise to revolutionize the office, they'll have to seep in discreetly, rather than slap users in the face. The machines, it seems, are ready. PCs have crossed the threshold of computing power needed for speech recognition. Now we need to work on the people. "Adoption barriers, generally speaking, are probably more on the human side of the equation than on the technology side," says Jock Mackinlay, a PARC user-interface researcher. "You don't want to disrupt a functioning office by throwing technology into it."

--With Marianne Kolbasuk McGee

Photo of Greiner by Bridget Barrett