Analysts frequently flood the computer-market press with predictions for the Next Big Thing. They're usually dead wrong - just ask all of the journalists who predicted the dominance of the NetPC.

In the last few years, these analysts have been talking up the WiFi "hotspot" market. A lot of companies launched services with which customers can pay to get 802.11b wireless internet access at a bunch of places around the country, such as coffee shops and airports.

There's only one problem with this: There's no market for it. They approached the problem backwards.

Free WiFi is great. It's quite nice to bring a laptop or PDA to a coffee shop and get some work done (or just goof off and waste time on the internet). It's an attraction to the hosting location. Given the choice between two otherwise-similar coffee shops, I'll choose the one that gives me free WiFi.

But I definitely won't pay extra for it.

Sometimes, coffee shops have a few soft chairs. When I choose a seat, I'll usually pick a soft chair if it's available. But I wouldn't pay for it. If the coffee shop decided to offer soft chairs only for an additional charge, I would disregard the soft chairs as an option - and therefore, I would consider that coffee shop to have only regular chairs. The soft-chair attraction would be lost, and I'd be inclined to go to a different coffee shop that let me sit in a soft chair for free.

For some reason, many locations have decided to contract their WiFi service out to a national provider that wants you to pay for it. If I want to use WiFi at Starbucks, for example, I have to pay T-Mobile $40 per month, $10 per day, or $6 per hour. I assume Starbucks gets a cut of this, and they get to put "WiFi hotspot" stickers on their doors to draw customers in. For Starbucks, this is simply another profit channel. Meanwhile, T-Mobile makes a killing, as their cost per location is almost zero: my apartment's 1.5 Mbit DSL connection is only $35 per month, and it came with a free wireless router. And I'm paying retail. If I bought a monthly subscription, I would completely subsidize the cost of one hotspot.

But what if I really, really needed internet access on my laptop when I was away from home?

Well, I'd have a few options. I could always get a dialup account for $10 per month that would work anywhere with a phone jack.

Or I could get a laptop cellular card for $119 and pay T-Mobile $30 per month (or $20 if you already have their voice service) for unlimited internet access, available everywhere covered by their cellular towers, whenever I want, without paying $4 for a cup of coffee.

I won't condemn T-Mobile or the other paid-hotspot vendors - they're trying to make some extra cash in a way that cost them almost nothing and requires almost no maintenance. It's the American dream.

I blame the coffee shops, hotels, bookstores, and other locations that host paid hotspots. They have completely missed the point. Free WiFi is a distinguishing feature of their establishments that can persuade customers to choose them over competitors. But discovering that the WiFi sticker on the door does not mean "free" is frustrating, and leaves customers with a negative attitude toward the establishment.

There is no "hotspot market". There is no good way for companies to make money by selling WiFi access. Free WiFi is an attraction to a location, and businesses should treat it like advertising: it costs money, but it will attract enough customers to cover its cost and provide some profit.