COMMENTARY-- Zigbee's getting ready to put wireless data everywhere, but an upstart company wants to nip in first. Will we get caught in the crossfire?

Back in 1917, when America was getting the hang of being a nation and flexing its youthful economic bada-bing, Vice President Thomas Marshall said with some feeling: "What this country needs is a good five-cent cigar." His dedication to unhealthy pleasures aside, it remains a good rule of thumb to look for popular stuff that can be made cheaply -- these are the things that change the world. What the world needs now is less carcinogenic but arguably more fun: a good one-dollar wireless chip. The Model-T Ford of the information age.

The Zigbee Alliance is well on the way to making this happen. This oddly named wireless consortium is dedicated to the idea of putting chips in light switches, toasters, keyboards and televisions: by concentrating on low cost, low power and appropriate range, it not only has a way whereby everything electric can hook into your digital world but the standards bodies are on its side. IEEE 802.15.4 may not be as surreally snappy as Zigbee, but they're part of each other and a good guarantee that what you get will work well with the mesh of standards already in place.

Only you can't get a Zigbee chip yet: mid 2003, they say. There is a dangerous chasm between waking up to an idea and having the bits to make it happen, and Cypress Semiconductor Corporation has rushed to fill it. For three dollars -- not a bad start -- you'll shortly be able to get its WirelessUSB part and build it into your gizmo. Low power, low cost, appropriate range: just not IEEE 802.15.4 (nor USB, it's fifty times too slow. Naughty.). But if it works, what's the problem?

Much the same thing happened a few years back, in the days when dial-up modems were exciting. Everyone wanted faster connections, but for a while things stagnated at around 33 kilobits a second. Physics said that this was about as good as you could get from a phone line. Then people realised that phone lines weren't exactly as the physics said, and that if you understood how the phone system worked you could squeeze out around two thirds more speed. The standards bodies started their labyrinthine grind, but who wanted to wait? So people got going with their own ideas.

The first mass-market modem chip that did these higher speeds came from a chip company called Rockwell -- easily the biggest supplier of modem chips around the world -- and they called their new standard K56Flex. Rockwell's most famous modem customer, Hayes, signed up and everyone waited for the international standards committees to make this good idea holy writ. But that would have been too easy: cue Rockwell's nemesis in the wacky world of datacomms, Lucent. They had their own pet modem maker, US Robotics, and neither of them wanted to cede bragging rights. "We're the biggest and most important people in the world of modems, and everyone's heard of us," said US Robotics. "We're very clever, and here's our 56K modem technology, X2. It won't work with K56Flex, because it's our market and we want it all." The protagonists puffed out their chests at each other and glared defiantly.

This didn't go down too well in the rapidly growing world of internet service providers, who had to decide which standard to back -- often ending up with two sets of modems on two dial-up numbers, and all the support hassles that that implied. The users had to find out which ISPs were running which type of 56k modem, then decide what modem to buy, then hope they didn't have to change providers. One daring modem manufacturer even tried to make a chip with both X2 and K56Flex -- and got threatened with legal action for its pains. I asked USR why they were persisting with this lunacy and was told, in a tone of voice normally reserved for particularly stupid children, "it's to increase consumer choice." Consumer choice said: wait for the proper V.90 standard. Widespread fast Internet access was delayed for a bit, and you probably don't have a US Robotics or a Hayes modem these days.

It's an annoyance when you've got to make a buying decision based not on quality or usefulness but on your guess as to which one will win. Betamax video recorders and the Commodore Amiga were both buried by standards wars. People got hit by the shrapnel. But with wireless, annoyance isn't the half of it.

The worst thing that can happen if you buy the wrong modem is that you don't get through to your ISP. But with wireless, if your new toy doesn't play well with others then not only will your joystick not control your videogame but your wireless data link may get befuddled by the strange noises it hears on its band. Zigbee's spent a long time making sure that won't happen with 802.11b or Bluetooth, both popular options that share its wavelength. Cypress' WirelessUSB also lives there -- and while Cypress says there are no interference issues, will that be the case in the future? Can one company exert influence on the standards bodies and keep up with the testing and interoperability issues? Is the company truly dedicated to standards-based computing, or has it spotted a gap in the market and decided to bypass that boring business of waiting for compatibility and just make some money while it's got the field to itself? A Model T is not much good if bits fall off and puncture the tyres of other road users.

Time will tell. But Cypress is playing a high stakes game with its low cost chips, and there may be repercussions for all.

Copyright © 2009 CBS Interactive, a CBS Company. All Rights Reserved.
ZDNET is a registered service mark of CBS Interactive. ZDNET Logo is a service mark of CBS Interactive.