The New York Times Circuits section did a fantastic job today in dissecting the hype behind 3G deployments, and making it clear to all and sundry that 3G is interesting, maybe useful, but isn't yet up to its own hype, and may only achieve part of its currently stated goals.

Katie Hafner wrote about the current generation of cell-data services, including the initial 3G versions. She concludes ably, But as is often the case with marketing carts that precede the technological horse, the bottom line is: Buy into 3G now and you are buying a promise. Wait another year or two and you are more likely to be buying the real thing.

David Pogue, Circuits's regular columnist, covers the bewildering pricing plans that include some of the variety of cell data services. Jeffrey Selingo offers accounts of the irritations with cell phone use and customer service.

Pogue also writes a brief column, much more fun and lose in tone, which is sent to subscribers to Circuits weekly email list. (Sign up here.) This week's column offered more insight into the alphabet soup of current and 3G data standards.

Both Hafner and Pogue (in his email column) fell into the trap of cell industry hype, however, without realizing it. That's how effective the marketing machine is. Pogue wrote, But that's only the first phase. By 2004, the big carriers hope to have true 3G (third generation) networks in place, capable of 384 Kbps data speeds -- and then a year or two after that, speeds of 2 or 3 megabits per second. You'll feel like you've got a cable modem or DSL on your cellphone. Hafner said, Forecasts vary, but as carriers upgrade software and equipment, they are likely to start advertising speeds up to 2 megabits per second by 2003. The actual speed will be closer to 500 kilobits per second at first.

Bravo on mitigating the hype, but both writers are following the lead of the industry in saying 2 Mbps, while that's a highly limited potential speed. It's an indoor speed. Both writers are accurate, but they left off the rest of the explanation, which the industry doesn't seem to want to talk in too much detail about. If you visit the part of the FCC's site devoted to 3G technology, the government speaks quite clearly on the topic. The three modes of 3G ultimately envisioned are: 144 kilobits/second or higher in high mobility (vehicular) traffic; 384 kilobits/second for pedestrian traffic; 2 Megabits/second or higher for indoor traffic.

Now think about that phrase: indoor traffic. Perhaps just as airports and public spaces like hotels and conference centers are finding that Wi-Fi might be useful to offer to their business customers, so, too, these venues might also invest in 3G indoor equipment. But most of the world won't. If you're a company with a building or a floor or a campus, you're going to need on-site cell telco equipment. That's called customer premises (or CUSTPREM) equipment in the Ma Bell terminology.

In the POTS (plain old telephone service) world, custprem equipment is a box at the demarcation (demarc) point where the customer runs his or her inside wiring from. You can often pay extra for the telco to run wire for you inside and terminate it, but their responsibility ends at the demarc point (unless you pay for inside wiring insurance, as they used to call it; they may not offer this in many areas any more).

The cell telcos aren't in the business of custprem equipment. They put up towers and transmitters, occasionally negotiating leases to put equipment at or near business campuses and run their fiber optic or other landlines back into their system to tie the phone service together.

To get 2 Mbps or higher speeds with 3G, companies will have to individually negotiate and lease equipment from cell telcos; or, if they're leasing building space, their building landlords will. But it won't be a given: it won't be everywhere.

3G frequencies already chosen in Europe and those that we are likely to use in the U.S. are in the sweet 2 GHz to 3 GHz range (a few bands drop below into the high 1 GHz band). These frequencies can only operate at distance with high power, which isn't and won't be approved for cell handsets.

This brings us full circle back to Wi-Fi and 802.11a and g. If, in two years, I have a full 54 Mbps/22 Mbps dual-band installation or some component thereof throughout my entire corporate campus - equipment that I own, have paid for, has present and future utility, and I control all aspects of - explain to me again why I would pay potentially tens to hundreds of thousands of dollars a year to bring the cell telcos into my facilities?

More likely, we'll see multiple band IP/GPRS/etc. phones. Give me 2.4 GHz in the office, in airports, at hotels, and I'll make VOIP or voice-over-IP calls (when 802.11e is ratified, we get the quality of service guarantees needed for VOIP). My VOIP phone will be tied back into my office's PBX or a private PBX a la Telesym's new IP phone, discussed yesterday.

But my magic phone will also have a software or hardware radio that tunes to other bands, including the seven or eight bands used for European and American 3G service. If I'm out of range of Wi-Fi or none is available, I can fall back into 3G. If I want to check my email while my cab takes me the 30 minutes from the airport to my hotel, I use 3G (maybe via a cable connecting my laptop directly to the phone rather than having a separate PC card, even, consolidating billing and accounts even further).

I've roamed far afield, but we're talking two to four years out. I want some good non-marketing proof that companies are stupid enough to pay for customer premises services when they can own and depreciate equipment themselves that offers superior connectivity and ties into their IT backends.

If You Made It This Far...More News

The FCC Approves Initial Ultrawideband (UWB) Standards: UWB is a very, very clever idea. Instead of using swaths of spectrum at low or high power,even rotating frequencies and jumping, you use incredibly short-duration high-power pulses across broad swaths of bandwidth that a similarly synchronized device interprets. You can use existing spectrum without (proponents hope testing will bear out) interference because existing devices rely on, would be affect or damaged by, or listen to a different pattern. By the time a pulse would pass, existing equipment wouldn't even tick over. Because of the broad amount of spectrum that could be reused, even many times in the same physical area, UWB might ultimately replace a number of existing technologies across a broad swath of consumer, scientific, medical, and military purposes. Here's a highly technical article on it from EE Times.