Municipal wireless always sounded like a great idea to me: A community gets together and puts up networking infrastructure for the benefit of its citizens. That’s got a certain all-American pioneer-spirit appeal, kind of like barn raising for the 21st century.

Unfortunately, reality isn’t quite that romantic. Many of the municipalities that eagerly embraced Wi-Fi early on are backing out. The latest casualty is Chicago, which announced ambitious plans in early 2006 and now says that it’s too expensive and not enough folks would use it. And San Francisco, Houston, St. Petersburg, Fla., and Alexandria and Arlington, Va., have come to similar conclusions.

What went wrong? The bottom line is that building and operating a wireless network is just plain harder — and more expensive — than it looks. (That chuckling you hear? Sprint/Nextel, AT&T/Cingular, Verizon and T-Mobile laughing in the background.) Many of the highest-profile municipal Wi-Fi meltdowns are the result of providers being unable to recoup the upfront cost of investment. That is, initial networks were set up “on spec,” with the provider footing the bill in the hope of attracting users downstream.

It’s now clear that this approach doesn’t work. Municipal Wi-Fi provider Earthlink, which is pulling out of several cities while struggling with a nearly $50 million loss so far this year, says the model is “simply unworkable.” Earthlink had planned to invest as much as $17 million in San Francisco, in the hopes of recouping costs by charging users $20 per month — but found it couldn’t break even on that plan.

And even when an experienced provider such as Earthlink or AT&T proposes to handle the operations, the peculiar complexity of operating in a municipal environment can lead to unexpected snags. Blogger David Strom highlights a case in St. Louis in which A&T was building a Wi-Fi network that relied on streetlight power lines for power to the access points. The only problem? Street lights only go on after dark, so the Wi-Fi network would only run at night. As Strom says, you can't make this stuff up.

This doesn’t necessarily mean municipal wireless is down for the count — only that the “build it and they will come” economic model is a total flop. One approach that continues to show promise relies on municipalities being willing to serve as “anchor tenants” on the network, providing a guaranteed annual revenue stream. For example, Minneapolis is planning to invest $1.25 million per year on its Wi-Fi network. Another approach — not my favorite option, but one that could work — would be to augment the funding with federal tax dollars. And of course, there’s Google’s approach, in which Google offers the service for free (and therefore, foots the bill).