Take your tablet
- 12 March, 2003 16:11
During a briefing on Tablet PC, one of Microsoft’s engineers copped to being outside the new platform’s target market. “I really need my notebook to be a desktop,” he said, dutifully tapping on the tiny screen with his stylus.
Amen to that.
Small, quiet, and efficient notebooks resonate with business users. They want to slip their PCs into valises and forget they’re carrying them at all. The rest of us trade chronically dislocated shoulders for the freedom to write code, run demanding applications and shade the occasional polygon in our hotel rooms.
Corporations see notebooks as sensible replacements for employee desktop PCs. Small and midsize businesses look to notebooks to connect their constantly mobile work forces. Vendors eye the business notebook market as a port in a storm. In this economy, few consumers are lining up to buy $9000 desktops, but companies routinely spend that much on one notebook with accessories.
As a result, mobile R&D, which had been in the doldrums, is firing up again.
Intel’s mobile Pentium 4 architecture removes old barriers that limit notebook performance. Microsoft’s Tablet PC OS draws more manufacturers into lightweight notebooks with a consistent design that’s certified for Windows XP. The wild card is the Apple PowerBook G4, a heavyweight that’s cruising toward most-wanted status not with raw performance, but with a unique mix of features that seems to strike just the right chord with business users.
Betting on performance
By 2000, notebook technology had hit a rut. Every Pentium III-based notebook with 256MB of RAM and a 14.1-inch diagonal screen was exactly like the one before it, except for an imperceptible CPU boost of 200MHz. Sales dropped off because vendors gave companies no compelling reasons to upgrade their notebooks.
I credit Intel, Microsoft, Apple, and Motorola with knocking portable computing out of its well-worn groove. Intel raised the ceiling on high-end notebook performance with the Pentium 4 mobile CPU (the P4-M). It saw no point in saddling a 2.2GHz processor with the 133MHz bus and memory standard in notebook designs. Intel mated the P4-M with a 400MHz I/O bus and 266MHz DDR (double data rate) memory. Manufacturers finally had something to sell that wasn’t warmed over from last year. AMD followed suit with a mobile version of its Athlon XP 1800+ CPU that also uses 266MHz DDR memory.
Performance always comes at a price though. In the case of the P4-M and Mobile Athlon XP that price is power consumption. Intel and AMD engineered their mobile CPUs to downshift into a low-power state when not actively computing. They have to — if these systems’ CPUs stayed locked at 2.2GHz (Intel) or 1.5GHz (AMD), batteries would last about 15 minutes. Trouble is, it isn’t easy for software to tell when it’s safe to switch the CPU to low-power mode. Windows 2000, Windows XP, and Linux are always computing in the background. Even when the CPU cuts back, other hungry components — particularly the display backlight and the hard disk motor — keep gnawing at the battery.
With all power-saving features enabled, my hefty yet lovable IBM ThinkPad T23 (Pentium III-M based) barely wrings two hours from its battery.
A colleague tells me his P4-M notebook gets about 90 minutes between charges. That’s pushing the definition of mobile.
Jot this down
Maybe the Tablet PC OS will fix the battery problem. Previous efforts to develop a mass market, touchscreen portable failed to catch on; notebook-based designs were too heavy to hold with one hand and Windows CE designs were too slow. The only way Microsoft would see a useful pen-based Windows PC was to get behind the wheel. The Tablet PC hardware specification standardises just enough of the hardware design to give vendors a more consistent target to shoot at — with the comfort of not going it alone in a new market.
One of the loose criterion Tablet PC sets out is battery life.
The spec says “up to eight hours”. I don’t believe Intel and AMD processor-based machines can approach that mark at Tablet PC’s suggested weight of 1.36 kilograms. Transmeta and VIA Technology both have Tablet PC-compatible, low-power microprocessors.
Systems built around these alternatives will give Tablet PCs a wider range of performance/efficiency options. My hunch is that second-generation Tablet PCs and similar non-tablet sub-notebooks will hit the five-hour mark. I think the market is sophisticated enough to accept a substantial performance hit in return for battery life.
For all notebook vendors, the no-brainer is connectivity. The Acer TravelMate 100 Tablet PC has what strikes me as a nearly ideal combination of connections: USB, FireWire, Ethernet, modem, VGA, and 802.11b wireless networking. I could wait another year for Bluetooth and integrated 54Mbps (802.11a) wireless.
Class by itself
The notebook market is being redefined, quite unexpectedly, by a system that looked like a niche player: the Apple PowerBook G4. On specifications, it’s big, heavy and slow. It doesn’t run Windows, and it doesn’t even have an Intel-compatible CPU. It’s a Mac, and only Mac people buy Macs. So PC vendors paid the PowerBook G4 no mind.
But competitors should take some notes. Motorola exempted the PowerBook’s PowerPC processor from the Intel/AMD megahertz wars, letting it meander under the 1GHz line until recently. Apple took an enormous chance selling 800MHz notebooks into a market yawning over 2GHz PCs.
Apple also gambled by shipping the Unix-derived Mac OS X after Dell and IBM failed to get business customers excited over notebook Linux.
Apple arrived at PowerBook G4’s design by making some daring guesses about what business users want. The inch-tall (when closed) PowerBook G4 has Gigabit Ethernet, a powerful ATI 3-D accelerator, a 15-inch diagonal display, and an internal CD burner. Thanks to its slower processor and aggressive power management in OS X, its battery lasts three-and-a-half to four hours per charge.
A similarly priced PC notebook can slaughter a PowerBook G4 in computing performance tests. But business users — along with Apple’s core of scientific, technical, and media customers — don’t run benchmarks. They run interactive applications, and for that purpose Apple’s flagship notebook performs admirably.
The same hardware that makes today’s desktops faster than last year’s servers is being adapted, in miniature, for notebooks.
Necessity is driving the process. In 2003, notebooks will sport Gigabit Ethernet, 54Mbps wireless, USB 2.0, and FireWire ports. Fast processors will be linked to ever-faster I/O and memory buses. In two to three years, users at the top of the performance spectrum should be able to add 64-bit PC and Mac notebooks to their shopping lists.
The solution to notebook heat and battery life problems is half in adjusting users’ perceptions and half in smarter hardware. Notebook users can’t expect desktop performance on battery power. Vendors must stop feeding that expectation. Future notebooks can take advantage of faster memory and I/O to do their work in short bursts, falling back to the lowest speed the user allows. That will strike a balance between speed and performance that should please all business notebook users.