Is the Pen Finally Mightier Than the Keyboard?

Microsoft says it has tablet PC software that works. Here's the inside story

For years, William H. Gates III and Warren E. Buffett have routinely mailed each other magazine articles that have caught their eye. They rip pieces out of the magazines, jot notes in the margins, and pop them in the mail. Gates anticipates the day when he won't have to mess with all that. With his new Tablet PC, he plans to call up articles from the Web, scrawl thoughts on the screen with a digital pen, and shoot it off to Buffett via e-mail. He's already using an early version of tablet software to send electronically annotated articles to Microsoft colleagues. "I have anticipated this for many, many years. And here it is," says Gates.

Few technical quests have consumed Gates as much as the notion of replacing pen and paper with an electronic tablet and stylus. When he brainstormed with Microsoft co-founder Paul Allen in the 1970s about what PCs would eventually do, he dreamed of machines that would process handwritten notes. He pushed Microsoft Corp. (MSFT ) into the business a decade ago, when several computer makers, including Apple Computer Inc. (AAPL ) with its Newton, made a run at pen-based computing. When that failed, Gates kept at it, funding research and waiting for the time when the electronic pen could become mightier than its inky ancestor.

Gates believes that time is now. Microsoft and its hardware partners have resolved many of the problems that dogged pen computing in the early 1990s--poor screen resolution, short battery life, and woefully inaccurate software for deciphering handwriting. On Nov. 7, the company will launch a version of its Windows PC operating system for the Tablet PC. Seven computer makers are offering this new breed of laptop PC with screens that double as electronic notepads. Some look just like a laptop--except that the screen swivels 180 degrees and folds down over the keyboard to convert to a writing surface. Others have screens that detach from the keyboard. The gizmos will retail from $1,800 to $2,500.

With advances in technology and Microsoft's marketing muscle, the Tablet PC stands a good chance of catching on. But it will take some time, particularly with tech spending so tight. Forrester Research Inc. analyst Frank E. Gillett believes only a few hundred thousand Tablet PCs will be sold the first year, partly because the user interface isn't as easy to navigate as it should be. But, Gillett says: "I'm bullish on this in the long run." Microsoft has spent $400 million on the technology and plans to spend $70 million more marketing the Tablet PC. Gates thinks at least 500,000 will sell in the first year.

Some early users are enthusiastic. In July, New York law firm Weil, Gotshal & Manges LLP handed out tablets to 20 attorneys in its Silicon Valley office. At first, the lawyers weren't quite sure what to do with them. Now, they won't give them up. They like being able to annotate electronic documents. "When these launch, we will buy only pen-based devices," says Weil Chief Information Officer James McGinnis.

Not everybody is so upbeat about the Tablet PC's prospects. Jerry Kaplan, a founder of pen pioneer GO Corp., dismisses Gates's new baby as science in search of customers. "There's no evident demand. This is technology push, not market pull," he says.

Moreover, Gates's revolution will have to start without two of the industry's most important soldiers: Dell Computer (DELL ) and IBM (IBM ). Dell President Kevin B. Rollins says it will wait to see how the market develops before jumping in. IBM is more circumspect. "We don't see significant growth opportunities," says Fran O'Sullivan, general manager of PC products and services at IBM. If that changes, IBM will get on board.

Gates believes they'll both jump in once they see the market develop. "If tablet is to do what I expect it to, you won't have anyone who makes portable computers who doesn't make tablets," Gates says. The tablet has become such a personal quest for Gates that his lieutenants feel added pressure to uphold his reputation. "I would be lying to you if I told you that I didn't wake up in the middle of the night in a cold sweat," says Alex Loeb, the Microsoft vice-president who leads the tablet effort.

It's no wonder Loeb's nights are tormented. The history of pen computing is littered with failures. Pioneers GO and Momenta defined Silicon Valley disasters in the early 1990s before anyone knew what a dot-com was. Microsoft's stab at the business, Windows for Pen, was launched in 1992, but made little headway. And the failure of the Newton's handwriting technology was so abject that Garry Trudeau mocked it in his Doonesbury comic strip.

The seeds for Microsoft's second run at the tablet were planted five years ago, when the company hired Dick Brass, who pioneered the electronic-dictionary business in the early 1980s. Microsoft wanted him to revive the pen-computing dream.

The first rough design for Microsoft's new tablet was made by a carpenter who worked on Brass's 55-foot Pilothouse Trawler. The wooden model had a flat base with a keyboard and a flat screen that could be carried separately. It cost Brass about $2,000. Brass proudly lugged the model into Gates's office for their first face-to-face meeting after he was hired in November, 1997. His new boss told him it was the kind of machine he had been dreaming about for years. But Gates gave him a budget for only six engineers, not the 20 Brass sought.

So Brass began assembling a team that would make up in brainpower what it lacked in numbers. He started by focusing on Microsoft's electronic-books effort. By mid-1999, he decided it was time to create the tablet, so he plucked two of the brightest scientists from Microsoft's vaunted research lab--Chuck Thacker and Butler Lampson. The pair had worked together in the early 1970s at the legendary Xerox Palo Alto Research Center developing the Alto, the world's first personal computer. Rounding out Microsoft's original tablet team was Bert Keely, who had headed tablet development at Silicon Graphics Inc. before joining Microsoft a year earlier.

To run the new group, Brass looked down the hall. A few offices away, Loeb was working on document imaging for Microsoft's Office group. Brass saw her as a spitfire who wouldn't be intimidated by the big brains of Thacker, Lampson, and Keely, or be cowed by pen computing's checkered history. It took some persuading. Brass wooed Loeb, a Georgian, over soul food from Dixie's BBQ in nearby Bellevue, Wash. "I would start by nibbling the cornbread. By the end, my face was greasy with ribs," Loeb says.

The first big bet the group made was to use Windows as the tablet's operating system. It seems like an obvious choice, but Windows is a power hog. That meant battery life, key to a mobile gadget such as the tablet, would be short. One crucial benefit outweighed the negatives: 10,000 applications run on Windows, so laptop users wouldn't have to give up anything by switching to a tablet.

At the same time, the team realized that they couldn't perfect handwriting recognition. Their software could accurately translate handwriting for only half of the users they tried it on. For the other 50%--forget it. So Microsoft emphasized storing handwritten notes.

Next, the tablet crew had to make a prototype to see if their ideas would work. Thacker knew that PC processors sucked up too much power. If the tablet couldn't run for more than an hour or two on its battery, users wouldn't take it to meetings or out of the office. One day in September, 1999, serendipity struck. While walking down the street in Palo Alto, Thacker bumped into Patrick Boyle, a colleague from his days at Digital Equipment Corp. Boyle, it turned out, worked for Transmeta Corp. (TMTA ), a startup that had a chip that used 70% less power than comparable processors. With the Transmeta chip, Thacker completed his designs by December.

What emerged was a gunmetal-gray gizmo that looked a lot like an Etch A Sketch. "It was pretty amazing for people to hold it in their hands and have Windows come up on that screen and just scribble a little bit," Gates recalls. "You'd say: `Boy, this thing looks like you could really do it."'

The coming-out party for the machine happened about 18 months after Brass got started, in November, 2000, at Comdex, the tech industry's annual hypefest. The team faced a mad dash to get ready for Gates's keynote. The software needed to be refined, new prototypes had to be made, and tablet marketing needed to take shape. "I felt like I was entering a vortex, and things were getting faster and faster and more compressed the closer we got to Comdex," Keely says.

Adding to the pressure, the new batch of tablet prototypes turned out to be flaky at best. The batteries weren't charging, and the screens could go blank if touched in the wrong place. Keely accidentally hit the power button during a practice run and shut a tablet down. Any kind of glitch during a high-profile product demonstration can spoil a technology's reputation. Charlton Lui, a lead software designer on the project, hid two spare tablets onstage so Keely could quickly grab one if the prototype he used crashed.

Luckily, that turned out to be unnecessary. Keely's machine worked flawlessly, and the techno-savvy Comdex crowd applauded as he strutted the tablet's stuff. Some of the coolest features included cutting handwritten words from sentences and having the entire sentence, and even paragraphs, slide together to fill the space. Lui, who has been working for years on handwriting technology, calls it one of the greatest moments of his life. "It was like I was floating in space, looking down on earth," he says.

The euphoria didn't survive field trials, though. In February, about 30 workers at three Minneapolis companies, including U.S. Bancorp, swapped their laptops for tablets. Microsoft dispatched teams of tablet staffers to follow them for several hours each day. The exercise was so consuming that ace usability expert Evan Feldman spent seven weeks in Minnesota, even missing his daughter's 7th birthday.

The findings were horrifying. The clever technology that wowed the Comdex audience drove users into fits. Workers routinely took notes that had no relation to the lines on the tablet page. They would scrawl diagonally, right next to notes that were horizontal. Sentence fragments merged crazily.

The technology that caused the trouble was among the most brilliant computer-science work on the project. Eventually, Feldman persuaded Loeb to dump it. "No magic is better than failed magic," Loeb says.

During summer, 2001, Loeb got even more nervous. Gates peppered her with e-mails about fundamental features such as the quality of the handwriting-recognition technology. Loeb feared he might pull back on the planned 2002 launch. She called a "come-to-Jesus" meeting in August in Gates's conference room, with two dozen execs from the Tablet PC, Office, and Windows groups to answer any questions Gates might have.

At the meeting, tensions surfaced between the Tablet PC and Office groups. Office executives wanted to focus on producing the next version of their applications, including Word and Excel, due two years later. But Loeb's team and Gates believed that the initial success of the tablet depended on having handwriting features in Office when the tablet launched. So Gates called for the tablet team to develop an "add-on pack," a collection of Office features designed specifically for the tablet. Then he gave the project the green light for 2002.

Gates even did some evangelizing of his own. In September, 2001, Hewlett-Packard Co. (HPQ ) agreed to buy Compaq Computer Corp. While both had been solid Microsoft allies for years, only Compaq had agreed to make a tablet, so Gates wanted to make sure the merged company went ahead full steam. On June 10, Gates met with HP boss Carleton S. Fiorina and No. 2 Michael Capellas to talk about the merger. "To one of those meetings, I brought my tablet, which was an Acer, and took notes and kept enthusing about how, `I don't think you're going to make enough tablets,"' Gates says with a smile. HP is a believer, and will be a major presence at the launch.

In August, Microsoft finished the software development for the tablet. The entire team partied the Thursday before Labor Day at one of the campus cafeterias. Keely's band, the Flying Other Brothers, rocked the crowd.

While the team put the finishing touches on its software, Gates made the tablet his indispensable tool around the office. It took a while for him to trust the machine. During the first weeks that he had the Acer test version, he still brought pads of paper along to meetings for note-taking along with his tablet. At one meeting in July, Microsoft CEO Steven A. Ballmer chided him. "`That tablet must not be very good yet,"' Gates recalls Ballmer saying. When Gates defended his tablet, Ballmer challenged him to leave his paper behind next time. After a few last fixes were made to the software, Gates did just that. Now he's betting the rest of the business world will follow suit.

By Jay Greene in Seattle, with Andrew Park in Dallas

    Before it's here, it's on the Bloomberg Terminal.