A Bluetooth failure

I had asked them to make the special trip to northeastern Massachusetts so that they could witness what I consider to be a total failure of the technology in a real world environment.

"The goal," says Bluetooth SIG marketing director Anders Eldund, "is to have a five-minute ready plan. Right now, the devices are not user friendly. Users can get most devices to work together, but only if they understand the concepts. For those who don't, it could be a lot easier to use. Users should be able to install, run, and connect in five minutes or less."

Because the user experience has varied so greatly from one Bluetooth-enabled device to the next, the Bluetooth SIG will be publishing implementation guides for each of the Bluetooth profiles (computer, PDA, phone, file exchange, headset, et al) which can be used by device manufacturers and vendors that want to ensure a more consistent outcome for their customers. Considering the state of the state of Bluetooth (especially in my experience), something like this is sorely needed. But will it be enough?

The first of these guides, for the computer and phone profiles, will be available in June. The remainder of profiles will be addressed in the months that follow.

Once Eldund and marketing programs manager Eric Schneider finished describing the implementation initiatives, we moved on to the main event. My goal was to demonstrate the frustration and confusion that one must go through in order to connect Bluetooth devices to each other. My guinea pigs were three Bluetooth-equipped notebook computers (an Apple PowerBook, a Toshiba Portégé, and a Dell D-Series) and one PDA (an HP 5450 equipped with Bluetooth and 802.11b radios).

These four devices' Bluetooth user interfaces had almost nothing in common. Even the two Windows-based notebooks had entirely different interfaces. Dell chose to go with Windows XP's Bluetooth management utilities, while the Toshiba had its own. The PowerBook user experience was in a world by itself because, well, it's from Apple. Of course, with a much smaller display, the PocketPC-based HP 5450 had its own unique look and feel.

Eldund and Schneider watched as I failed repeatedly to connect from one device to another. None of the attempts to connect from the PDA to any of the notebooks were successful. In fact, the only successful file transfer that I was able to complete was a Dell-initiated one from the PowerBook. Eldund even tried connecting from his Compaq iPaq to my PowerBook with no luck. The attempts were met with confusing dialog boxes and error messages.

Even the language in the dialogs was far too technical for most mortals. Instead of something as simple as "Download a music file from another computer," I would get a list of services to engage, like "OBEX File Exchange." I don't know or care what OBEX is, and neither should any user. The dialog boxes were riddled with confusing instructions, extraneous text, mangled device names, including instructions to launch another application if I wanted to do something else related to Bluetooth. (Why not have it all in one place?) The user experience of configuring, connecting, and engaging in some Bluetooth-related task was entirely different from one device to the next. In fact, about the only thing that was consistent across the all the devices is that Bluetooth simply didn't work.

Eldund explained that it's exactly this sort of problem that the SIG's forthcoming implementation guides are supposed to address. When I confessed that I planned to write about this sub-standard experience, the SIG representatives revealed that consortiums like Bluetooth SIG are relatively powerless when it comes to cleaning up messes like this one. The consortium can make recommendations, but it's up to the vendors to take those recommendations into consideration. I was actually encouraged to write this story because it can only legitimize many of the consortium's recommendations.

So, in addition to all of the recommendations the Bluetooth SIG is working on, here's one more. Instead of coming up with a implementation guide, and then having each vendor build from scratch it's own implementation in hopes that they all work the same, I have a better idea. The Bluetooth SIG should join the Java Community Process and create a standard API so that a Java Virtual Machine can access the Bluetooth hardware.

Then, the Bluetooth SIG should develop a single Java-based application, one for each of the platforms (phones, PCs, PDAs, etc.) and distribute that application to the 2,000 vendors that are members of the Bluetooth SIG. Then, from one PC to the next, from one PDA to the next, and from one phone to the next, the application for getting something done with Bluetooth would look the same. Of course, I don't expect Microsoft to jump at this idea and include it with their operating systems, but, that doesn't prevent the ultimate system vendor from including it before those systems go out the door.

There you have it: the wild, wacky, Bluetooth truth. I wish I was lying. But Eldund and Schneider saw it with their own eyes. Hopefully, you won't have to.