Vol 16, Issue 8

February 25, 2002

Competing for the Palmtop

By Max Baron

If the PC word processor has replaced the typewriter, and the spreadsheet has replaced tabulation sheets, then the palmtop has effectively replaced the little calendar book. The definition of a palmtop is vague. Some vendors call their products "palmtops"; others use such descriptors as "handheld" or "personal digital assistant" (PDA). Instead of trying to define what is and what is not a palmtop, consider the Palm Pilot and the similar market-visible devices that compete with it. I believe the palmtop's evolution has reached a turning point, brought about by recent microprocessor introductions and company acquisitions.

Born in 1996 at Palm, Inc., one year after the company's acquisition by U.S. Robotics Corp., the Palm Pilot became popular and can be credited with the resurgence of handheld computing. Light and useful, the Palm Pilot provided a daily and monthly calendar—a good place to enter appointments, expenses, addresses, telephone numbers, and short notes. All you had to do was fork out a few hundred dollars and learn to write in a somewhat new set of characters—and you could throw out your five-dollar paper calendar.

The Palm Pilot was originally powered by a 16MHz Motorola 68K-based DragonBall chip that, after two frequency jumps, attained today's breathtaking speed of 33MHz—enough, however, to process simple handwriting and data entered by stylus. A handful of games could be played on the palmtop's screen, the few available pixels reminders of the early days of Atari.

Then came improvements. Screen contrast was improved, so you could actually see what you were writing under less-than-intense light. This improvement generated more revenue than expected for the manufacturer, because the Palm now cleverly came in a slim, attractive metal case and featured built-in rechargeable batteries. (Discard system when batteries fail?)

The new palmtop's price was close to $500, and it provided roughly the same features as the less-expensive one, its sales a marketing success that should be studied in universities. Palm and licensees began to compete with each other in transforming the palmtop into a piece of jewelry akin to the digital watch. Whose product would look better, be slimmer and lighter, and have more accessories to pull out of leather bags to elicit "Ooh!"s and "Aah!"s from envious friends and relatives?

In April 2000, Compaq, undaunted by its former failure with Win CE PDAs, introduced the iPAQ, a handheld (Compaq's term for a palmtop) that used Microsoft's Pocket PC OS to deliver a sophisticated set of applications, quality graphics, generous backlighting, device and memory expandability, connectivity, and, most important, compatibility with the higher-end Microsoft Office programs. The processor that made it all possible was Intel's StrongARM, designed at Digital and running at 206MHz. The iPAQ drained batteries faster, and it was heavier than a Palm, but comparing it with the first Palm Pilot was like comparing a PC to an Atari box. Compaq and Microsoft had defined a new generation of performance, luxury, and expensive palmtops. Casio, HP, and others offered similar devices.

Palmtop features and performance are becoming the next competitive frontier. A high-performance processor must have a place to store large files of instructions and data. The Pocket PC group had StrongARM performance and had already incorporated large internal memory and expansion slots. The Palm OS team needed higher performance and increased memory resources.

Motorola came to the rescue. The announcement of its ARM 9–cored DragonBall MX at 200MHz underscored the news of its architectural license purchase from ARM. DragonBall, StrongARM, and TI's TMS320VC5470 (DSP and ARM on one chip) can make ARM, the architecture, the target of applications and operating systems for handheld devices. While we were still wondering about what we'd see on the palmtop—a little house, a rectangle marked "Start," or a tiny penguin—Intel announced new StrongARM technology: the XScale-based PXA250, running at up to 400MHz, and the less-expensive PXA210.

Intel's announcement came just days after AMD's confirmation that it has acquired Alchemy Semiconductors. AMD is building a group of more than 300 people to focus on high performance at low power, beginning with the 500MHz, 32-bit MIPS-based Au1000. The MIPS low-power technology is no stranger to palmtops: the recent BE-300 Cassiopeia boasts a 64-bit MIPS-architecture NEC VR4131 chip that can run at up to 250MHz (230mW). The Cassiopeia clocks the chip at 166MHz to run Microsoft's Win CE 3.0 OS on a high-quality color palmtop that, on good sale days, goes for $149.

Microprocessor frequencies of 400–500MHz, soon to be seen in palmtops, were the highest ones could buy in Pentium PCs three years ago. The recommended PC internal memory of those times, 64MB, has already been replicated in Pocket PC 2002 palmtops. Pricewise and technologywise, the palmtop is still far from enabled to fully use a 500MHz processor in the same way the PC did. Handhelds' disk storage capacity still has a long way to go. Battery capacity needs development (fuel cells?). As for connectivity: At the time of this writing, wireless services, at $39.99 per month, annually equal the price of one new top-of-the-line palmtop.

At the high end, we should expect pricey new high-performance palmtops—some delivered in expensive packages. At the low end, decreasing prices can bring some palmtops to well below the $100 pricetag. At CompUSA recently, I saw two 11-year-old kids discussing the merits of different styli for their Palms. Handspring is offering a joystick attachment for its palmtops. Will there soon be a palmtop equivalent of the digital plastic wristwatch?

Introductions of high-performance microprocessors, powerful operating systems and application software will continue to provide high-performance luxury palmtops for those who like them and can afford them. Competition will lower the prices of the other palmtops and the microprocessors inside them. The palmtop can come within a range that schoolchildren can afford. It can become their game and music platform—and a new tool for learning.