Father of an industry

DN Staff

March 6, 2000

16 Min Read
Father of an industry

During a rare moment of quiet in his hectic day, emergency room physician Steven Epstein pauses to review notes on a specific medical procedure on his Palm hand-held computer.

Meanwhile, miles away, Quality Assurance Group Leader Steve Isenberg, of Nortel Networks, checks his Palm IIIx to look up the phone number for booking a business flight.

And, hard at work in his office at the Massachusetts Institute of Technology, research engineer Bob Childs hears the beeper from his Palm Vx, reminding him that he has 10 minutes to get to a meeting across campus.

All three, like millions of other professionals, business people, and general consumers across the country, have come to depend on these pocket-size devices to help them manage their lives.

They're indispensable, says Epstein.

They're versatile, adds Isenberg, who even keeps track of phases of the moon with his.

"They're my lifeline," says Childs, who admits to having missed many a meeting before he got his.

But they almost were a product category that died an early death until Electrical Engineer Jeff Hawkins, now chairman of Handspring Inc., had a revelation.

It was 1994 and the then-37-year-old Hawkins, who had started Palm Computing after stints at GRiD Systems and Intel, had just experienced failure with his company's first product, the Zoomer. A cumbersome, if feature-laden, hand-held computer, it suffered the same fate as the Apple Newton, which beat it to market--and infamy--by two months. No one wanted either because they were too bulky, slow, awkward, and quirky.

Instead of trying to jam the product down customers' throats with a snazzier upgraded version, like Apple and, later, Go Corp., Sharp, and others did with their products, Hawkins thought, "Why not ask the customers what they really wanted?"

Their answer: Don't try to replace our desktop computers; just replace our pocket and desk calendars.

Eureka! "I realized my competition was paper, not computers," Hawkins says.

And so he began a two-year design effort to develop a simple, no-frills, compact device that hit the market in 1996 as the Pilot. All it did was hold addresses, phone numbers, a calendar and a to-do list the user could write on. But simple as it was, it quickly became the fastest-selling computer product ever, according to the business magazine, Fast Company.

Brand new industry. Since then, a slew of other companies have entered the hand-held computing market, which has spawned at least four specialist magazines, hundreds of peripherals manufacturers, and formal user groups filled with customers who can't stop talking about--and using--their hand-helds. Palm's installed base is about six million, and Palm has practically become a generic name for that class of product.

"Essentially, Hawkins created an industry with the Pilot," says Michael Steinberg, a technical consultant and president of the New England Palm Users Group.

"He made a courageous move" going smaller and simpler, says Pen Computing magazine's Palm Editor Shawn Barnett. "Everyone up to then was dissatisfied with hand-held computers and ready to go back to their PCs."

Hawkins followed the Pilot with several enhancements, beginning with the PalmPilot and continuing to the Palm VII. Today, he works out of Handspring Inc., which he started in 1998 after failing to convince 3Com, parent of Palm Computing, to spin off the Palm unit and invest more in the development of the product. 3Com has since done just that.

Sounding part proud father, part missionary, the high-energy Hawkins talks passionately about the future of hand-held computers, and about his latest product, the Visor. It uses the Palm operating system, but enables users to plug in a variety of small devices for playing music, computer games, and other activities. What enables the plug-ins is a 68-pin expansion slot within the device called the Springboard. Pen Computing's Barnett says it's the Springboard that makes the Visor an exciting and revolutionary product. "The Springboard slot gives peripheral designers the gateway they want to the Palm OS (operating system) marketplace," he says.

Which, curiously enough, has been one of Hawkins' goals since the beginning. Even while scrapping features from the original Palm product to make it simpler and smaller, he always wanted to make the Palm a platform so others could write software for it. "I know users want to extend their Palms, and that's why we introduced the Springboard expansion slot," he says.

"Jeff understands the user experience and instinctively knows what's important to them," says Peter Skillman, senior product designer at Handspring who first worked with Hawkins as a consultant for IDEO, the engineering firm that teamed with him on several of the Palm products. "He has a real empathy for customers."

It's as much a part of his makeup as the informal, affable demeanor, say associates. And that empathy shows itself in ways that might seem odd to others.

Donna Dubinsky, president of Handspring and former president of Palm Computing, remembers people watching Hawkins pretending he was taking notes during meetings on the original Palm using a pencil and a rough wood prototype of the product. "They couldn't figure out what he was doing," she says. "I told them he was inventing."

And she was right. He had determined that the handwriting-recognition software other products were using was too complex and slowed the devices down, so he developed a new software called Graffiti that required users to learn a simplified alphabet the computer would quickly recognize. He was practicing with it at those meetings to see how comfortable and easy it would be. Graffiti became one of the differentiators of the product.

Design control. That wood model--and many others that followed it--have helped anchor Hawkins' creative thinking to the needs of users. And, the wood models have helped him make sure that the various design iterations don't stray from his original ideas.

Dennis Boyle, studio leader and one of the founders of IDEO, recalls how Hawkins used his wood model of the original Palm to control the size of the end product and resist what Boyle calls "creeping elegance."

"He did a stackup to see how thick the product would have to be to hold two triple-A batteries, which are about 10 mm in diameter," Boyle says. "Then he carried it around in his pocket just as he imagined users would do." When the engineering team would show him designs that were thicker than his shirt-pocket-size model, he would send them back to their workstations.

Hawkins had determined that size was critical with the Palm. But controlling the size was practically a full-time job. "He battled on every millimeter," Dubinsky recalls.

But not in a hostile way. Among his strengths, says Boyle, is his general good nature. "He is a personable, nice guy who is approachable and listens." But, Boyle adds, you can't convince him with half-formed ideas.

The original Palm team found that out early. To save space, Hawkins wanted to put the triple-A batteries under the digitizer. Many competing products had them at the bottom of the unit where the buttons are, but that would have increased the thickness in the Palm.

He pushed the team to use thinner plastics. And, he insisted on replacing the backup battery with a super capacitor to save space. Every previous product had a backup battery that took over when the main battery was being changed. Replacing the backup with the super capacitor saved space and reduced complexity, but it was controversial because it would only power the Pilot for a minute or two while changing batteries vs. hours for a backup battery. "It was the right call and we never regretted it," he says.

"There's a natural tendency to complicate things and make them more complex, and a reluctance to take risks," Hawkins says. "But there is no shame in failure, as long as your ratio of success to failure is positive."

He inherited that attitude and his inventiveness from his father, an engineer who was always inventing new products like an air-cushion boat (which the teenaged Hawkins crashed into a Manhattan bridge over the Harlem River, tying up rail traffic for an afternoon), and technology for recognizing dolphin speech.

"Dad did a lot of interesting things like that, and I sometimes worked with him," he says. "That's how I learned about fasteners, materials, and how to use tools." The knowledge has come in handy.

"I build computers, which requires knowledge of electronics and software," says Hawkins, who got his electrical engineering degree from Cornell University. "But they're also very mechanical, and so knowledge of mechanical engineering is important."

For example, his goal for the Palm III was to make it more rugged, and that was more a mechanical than an electrical challenge. He added strength through design by, among other things, using four screws instead of two for the housing, and using additional tabs in the side of the housing.

He faced a different challenge with the Palm V. Microsoft at the time was entering the market with a product based on their Windows CE platform. Hawkins decided that rather than matching Microsoft feature for feature, he would do something he believed Microsoft couldn't do: make a beautiful product. Glue, rather than screws, met the aesthetic goals, but the engineering team balked, saying that using glue would be difficult. But Hawkins had worked with glue before, unlike many on his team, and knew an adhesive could be made to work. 3M's VHB did.

He also made the V even thinner than its predecessors. He accomplished that in part by getting rid of the separate internal frame the previous Palm products had and incorporating a mid-frame. The design incorporates two thin aluminum face plates glued together to form the backbone of the product. The mid-frame also serves as a style element, providing a dark band all the way around the product.

"Jeff has the ability to zero in on the minutest details, like plating, or how much use a snap fit can take before it breaks," says IDEO's Boyle. "He can see the big picture and the small details simultaneously, which is rare."

And, adds Dubinsky, his business partner and president of Handspring, he has good business instincts even if he is uninterested in day-to-day business operations. "For example," she says, "if I explain an accounting problem, he will add value to the discussion. He asks questions others haven't thought about, and he prioritizes well."

Keep it simple. The biggest mechanical problem in design of the Visor was making the Springboard slot and connector low-cost and reliable. He accomplished both by using the same pin-block design used on standard PCMCIA cards. He also changed the keying method to prevent accidental insertion of PCMCIA cards.

There was also the problem of the force needed to pull the Springboard connector out of the Visor: It was so high that, when engineers passed a prototype around a room, the only person who could remove the connector was a woman who was a triathlete. "A totally engineering-centric person would have lowered the force, but that would have weakened the connection," Skillman says. Instead, Hawkins had the team design a dent in the back big enough for a thumb to provide leverage for the user to disconnect the Springboard.

A wild idea? Perhaps, but, says Skillman, Hawkins is an engineer who encourages wild ideas. And that tendency ties in very nicely with one of his extracurricular activities: the study of human intelligence. Actually, to call his interest an extracurricular activity downplays its importance to him.

"I've been interested in the human brain for years," he says. In fact, he left the business world in the mid 1980s to pursue a doctorate in biophysics, but left the program when his faculty advisor wouldn't let him do his thesis on a theory of how the brain works.

He did gain some insights into the brain, however, that have helped him in his work on hand-held computers. Those insights involve the concepts of understanding and auto-associative memory, which can work with in- complete or erroneous data. He discovered how to apply auto-associative memory to data that varies with time, and used the discovery as the basis for the mathematics behind the Graffiti handwriting-recognition software in the original Palm.

Eventually, he hopes to fund research into the theoretical aspects of distributed memory.

"I want to make my mark in business so I can ultimately have a forum for my work on the brain," he says. "I want to change the world in a small way with my work."

He has already done that with hand-held computers. We await with anticipation the changes he'll make with his studies of the human brain.

The right connections

Pen Computing magazine has called the Springboard platform the element that makes the new Visor hand-held computer "exciting and revolutionary." It's a jumping off point that will make the Visor more than an organizer, the magazine says.

The platform is an open-face expansion slot that Jeff Hawkins and the rest of the Handspring Inc. engineers designed to easily fit together with a broad range of accessories called Springboard expansion modules. Here are some of its features:

  • Open-faced slot. The open face is on the back and top of the Visor and accommodates peripherals that extend out the back and beyond the top. Rails on both sides of the slot enable Springboard modules to slide in easily, yet not fall out if the unit drops on the ground.

  • Automatic install and de-install. Handspring engineers say that installing software in desktop computers often can corrupt certain files to an extent that only an expert can remedy. So they designed Springboard to self install and de-install. In fact, the hand-held computer can be on or off--even running an application--when a user removes the module, and no damage results. The reason: Special software handles all installation chores, detects when a user removes a Springboard module, and then de-installs everything related to the module.

  • Flexible power. All Springboard modules can count on 100 mA of current at 3.3V. If the user needs more, he can incorporate his own battery in the open back of the Springboard slot.

Engineering basics beat problems Product design and development always comes down to solving problems. But the solutions don't always have to involve revolutionary new components or materials. Here are three problems Jeff Hawkins and his team faced with hand-held computers and how they solved them with good old-fashioned engineering thinking: - Ruggedness. When it came time to design the Palm III, Hawkins and his team decided to concentrate on improving its ruggedness. After some study, they found that most breakage was occurring in a corner of the digitizer, where a right angle in the metal frame protected against bending. Solution: They used four 1.7-mm self-tapping screws instead of the previous two in the frame, then inserted tabs to strengthen the plastic housing. On the Palm V, they made the infrared window part of a new mid-frame, which they made from an infrared-transmissive material. To have used a separate infrared window would have required holes in the plastic, which would have weakened the frame. Molding it instead into the mid-frame made it stronger. - Footprint. After realizing that users wanted their hand-helds to replace their desktop calendars, not their PCs, Hawkins concentrated on simplicity and size. He stripped away unnecessary features, and looked for ways to cut down the size of the device. First step: He built a wooden model thick enough to hold two AAA batteries and narrow enough to fit in his shirt pocket. He does that for every product he designs. Then, he pressed engineers to find materials for the display and housing thin enough to accommodate that size and still work. On the Palm V his goal was to make a fashion statement, which required an even-smaller unit. The engineering team said it would be difficult to use a plastic for the housing that was less than a millimeter thick, so he suggested metal. That's a common material for housings today, but not in 1997 when he was working on the Palm V. "Flat-sheet aluminum was used in PCMCIA cards, and we knew we could get it in 0.5-mm thicknesses, but it had no curves," he says. To avoid a boxy look, the team found a molder in Japan who could put curves in the aluminum. Since there was no room for an internal frame, engineers moved the frame to the outside of the product. They specified that two aluminum faceplates be glued together, not fastened with screws as in previous products, to form a rigid structure that is the backbone of the product. It serves as a style element, providing a dark band all the way around the product, and holds the PCB and display. To make the Palm V narrower, engineers put channels in both sides of the housing to hold the pen. "The pen actually becomes part of the outside of the product, which saved about a millimeter," Hawkins says. The Springboard slot on the new Visor presented another sizing problem. The slot actually takes away available space for the electronics in the product. Hawkins and his engineers relocated the batteries under the buttons, but that required thinner buttons. Earlier Palm products used elastomeric switches that required travel of 1 mm with a button height of 2 or 3 mm. Snaptronics, Inc. provided Hawkins with 0.5-mm metal snap-dome switches. - Component costs. Controlling component costs involves more than simply finding the least expensive vendor. Hawkins and his team did their fair share of vendor searching, but they spent more time thinking through their designs and trying to incorporate standard products. For example, with the Visor, the team looked at several design options for the Springboard connector. "Connectors are always bugaboos in electronic design," Hawkins says. "We were starting from scratch, so we could have gone with any connector design." To save costs and make it easier for third-party vendors to provide peripherals for the Visor, Hawkins instead decided to use the standard 68-pin PCMCIA connector design, already familiar to peripheral manufacturers and available from several vendors. To avoid the problem of users plugging a PCMCIA card into the Springboard slot, Hawkins specified that the key on the side of the connector be turned upside down so standard cards wouldn't fit in.
Sign up for the Design News Daily newsletter.

You May Also Like