Don't bet on it.

There are enormous engineering challenges stemming from the move to broadband. I've heard estimates that a 3G handset running wideband CDMA will need two to three orders of magnitude more DSP horsepower than a current PCS handset just to deliver data to the CPU. Then there's getting all the necessary data-processing power into a small, battery-operated package. The 3G hand-set will have to deal not just with compressed voice, but with Web pages, streaming video and all manner of demanding data types. So there will be a lot more applications processing as well, and a big display to run-all with decent battery life.

As for software, given the range of stuff 3G devices are supposed to do, they would appear to need a mix of Windows, Netscape and some really hot multimedia apps, all running under a real-time kernel, from ROM, in a palm-top package.

But I don't think those will be the problems. Based on my experience with everybody else's digital PCS phones, I think the problem will come from another quarter. The technology will work, but the operating companies will mess it up.

A couple of years ago, EE Times interviewed systems engineers working on then-new CDMA technology. They weren't worried about building base stations and handsets. They were concerned about touchy deployment issues-the topology of the cells, in particular. They feared greedy system operators would mess it up. They feared that in the race to cram even more active calls into the available bandwidth, vendors would push voice compression beyond reasonable limits.

Well, guess what? I can always tell when I'm conversing with a victim of the new digital handsets. The voice quality is appalling-frequently I can't identify folks I know well, and on occasion I can't even identify gender. Interchannel interference is at least as bad as it was in crowded AMPS service areas. Digital service even brings a new twist: Sometimes you simply get switched into another call, instead of mixed in with it.

The percentage of calls that are completed instead of terminated by problems is absurd. In some areas it's normal to take two or three phone calls to get one simple message through. Frequently that message has to be "I'll call you later."

What will happen when system operators get their hands on 3G technology? Or, to put it a different way, what will happen when I try to load a Web page onto my 3G communicator and I have to reconnect six times before I get it? How likely am I to use the personal-productivity applications on a portal site if transmission problems reduce my effective data rate to below my typing speed?

It's bad enough that the digital revolution made voice service worse instead of better. But there's a very real risk that deployment errors that are beyond the control of the technical community will transform this new rollout from universal connectivity to an exercise in futility.