I really like Mike Langberg's take on municipal wireless hype: Langberg is a long-time tech journalist, and filed this column for the San Jose Mercury News based in part on visiting Esme Vos's MuniWireless 2006: Silicon Valley conference this week. He points out the weakness in current discussions of wireless: last year's poster child is this year's red-headed stepchild. Chaska was last year's darling; St. Cloud, Florida, this year's.
Now the problem isn't with a particular city or vendor. Rather, and let me be the first to admit this with complete frankness, those of us who write about municipal wireless have to take the word of both the people building the networks and the people castigating them. When I say "word," I don't mean an uncritical evaluation and repetition of statements made by those parties. Rather, I can't drive the streets and walk into people's homes with signal detectors at different times of day over a few weeks for every city I write about. Or any city I write about.
Hotspot service you can spot check. Metro-scale service is a statistical problem. If 50 percent of the city gets perfect access everywhere and 50 percent can barely hold onto a signal, that's not a functional network. But what if 90 percent gets a great signal and five percent receives an unusable one? It's impossible for a journalist to know.
Langberg hits the nail on the head when he quotes folks in the industry noting how difficult it is to get straight answers. Here's him quoting Google's metro wireless head Chris Sacca: Equipment makers trash each other instead of offering hard performance data, Sacca complained, then make ``outrageous and exaggerated claims'' about their own product
I had a line in a draft of my Wi-Pie in the Sky article that appeared in The Economist in March about how every vendor I spoke with first complained that Tropos's gear couldn't possibly work as advertised, and then described how nearly every other metro-scale competitor's equipment also wouldn't work. Of course, two firms (BelAir and Motorola) make single-radio equipment similar to Tropos, which defeats their arguments a bit, I fear.
Metro-scale Wi-Fi can work and it will work. But the scale and complexity of projects combined with the low density of Wi-Fi nodes appears out of whack to my analysis. The only way to know? When projects are deployed and third parties produce rigid engineering analyses.
Update: Of course, as soon as I write these words, one reporter proves me very partly wrong. Mike Rogoway of The Oregonian traveled around MetroFi's Bay Area coverage range for an idiosyncratic taste test. He found extremely variable performance. But the service performed admirably in many places. And, Rogoway doesn't note, the Bay Area installations in the South Bay were not built to a city's RFP. These were privately deployed with no city contract. Portland will have performance measurements to meet, and MetroFi might have to deploy a greater density of equipment to meet those marks.