No mobile and wireless technology has been surrounded by more hype than Bluetooth. Often described as "a wireless wire," Bluetooth is a form of short-range wireless connectivity designed for point-to-point connection of local devices. Bluetooth demonstrations often include wireless printing from Personal Digital Assistants (PDAs) and wireless connections between belt-clipped cell phones and headsets. An extraordinary set of companies backs Bluetooth, including 3Com, Ericsson, IBM, Intel, Lucent, Microsoft, Motorola, Nokia, and Toshiba. For more than 2 years, backers have predicted a Bluetooth revolution in mobile computing.

Well, if you see any sign of that revolution, please send me an email message! I seem to be missing it!

Let me first praise Bluetooth as a neat technology—but one that's seriously oversold. Microsoft once believed Bluetooth could save from oblivion the palm-sized PC (the predecessor to today's Pocket PC). Motorola's Software Products Division is urging OEMs to build Bluetooth support into the next generation of notebook PCs, and not a week passes without a new press release touting another Bluetooth device or application.

But the timing of the Bluetooth revolution keeps getting pushed back. What went wrong?

First, Bluetooth specifications took time to stabilize. Late in 2000, version 1.0b moved to version 1.1. Now work proceeds on version 2.0, which promises higher performance and longer range. That's all good—but we're talking about a device protocol that can't be used on an installed base without established standards.

Second, I witnessed a serious usability problem at last fall's COMDEX trade show. I attended a demonstration of a Bluetooth-based wireless printing solution used to send a document from a Handspring Visor with a prototype Bluetooth Handspring add-on module to a printer plugged into a Bluetooth hub—in the Bluetooth pavilion.

The desired printer responded—along with dozens of other devices within Bluetooth's working range (tens of feet)! Using the Handspring Visor's screen, Bluetooth built a long list of device identifiers composed of apparently random letters and numbers. The demonstrator had to walk to the target printer, compare its tag number with the list, and select the correct identifier. Only then did the printer produce the correct document.

Now that's hardly a fair test, because only at events such as COMDEX do we find several dozen Bluetooth devices communicating at once in a relatively small space. And I know that various companies are preparing browsing solutions that will present friendly device names such as modern networks use. You can finesse Bluetooth protocol specifications by implementing the protocol in Flash memory rather than ROM (so you can re-flash as the specification changes). But the cold fact remains: Bluetooth lacks the stability—as of now—to become a major force in wireless networking.

Moreover, Bluetooth must now compete with 802.11 wireless Ethernet (aka Wi-Fi), a much more stable technology. In fact, Starbucks recently announced plans to provide 802.11-based wireless Internet access in some stores. That's a powerful statement about 802.11's increasing popularity as a de facto standard for wireless networking. Compared to Bluetooth, 802.11 offers longer range and easier integration with existing Ethernet-based LANs.

Yet Bluetooth makes far more sense than 802.11 in some applications, such as wireless cell-phone headsets. But for printing or more general uses, I'm not so sure. What do *you* think?