14. Testing Procedure

We used the test plan for sub-notebooks as the basis for our tablet PC testing procedure. After all, tablets are technically similar to sub-notebooks, and even the requirements are the same: high ergonomics before breakneck system speed. For this reason, the lab experts rank ergonomics disproportionately high, at 40 percent.

Other factors in the ergonomics ranking are parameters that are standard for notebook tests, such as battery life, for example. The two top tablets proved to be especially long-lived, the machine from HP with a very good four hours and four minutes, and the Fujitsu model with three hours and 37 minutes. Battery life is determined with Business Winstone 2002 Battery Mark. To measure comparable runtimes, the lab experts set the display brightness for all candidates at 100 candelas per square meter and disable any existing W-LAN or Bluetooth modules. The benchmark simulates the work of a user who intersperses usage times with breaks of different lengths.

Also important for the ergonomics score is the weight, which counts for 30 percent of the rating. The lab experts weigh the devices including all optical drives, but without the power supply unit. The latter usually remains at home, so the measured weight (without the external power supply) more closely approaches what it is in practice.

To take the new tablet PC device category into account in the test, experts review several additional Tablet-specific aspects as well. These include removable keyboard, slip-proof underside (important for holding it conveniently on one arm while working) and arrangement of the connectors on the case. As far as the latter is concerned, it is important to check that the most often used connectors, such as the power input and Ethernet port, are arranged in such a way that the cable will not keep the user from working either in portrait or landscape format.

Performance Test Bench

To rate the system performance of the tablet PCs, Content Creation Winstone (CCWS) 2002, Version 1.0, is used, among other things. CCWS determines the overall performance of the device. Only when all components (processor, graphics chip, hard drive and memory) are fast will the device achieve a high rating. What stands out among the tablet PCs is that the two devices with Transmeta CPUs (from HP and Paceblade) are far behind their rivals in CCWS, with 8.4 and 7.7 dots, respectively. The combination of chipset and CPU is not nearly as powerful as the Intel components.

Pure CPU performance is tested with the Povray raytracing software. The calculated values confirm the CCWS results, as here too the Transmeta devices produce considerably worse results than their competitors.

Serving as the Povray demo picture is "Chess2," which is calculated in a 640 x 480-pixel resolution and with the anti-aliasing setting "0.3." The lab experts convert the number of seconds into points to better compare them with the other benchmarks.