London is full of interesting archaeology from dead civilisations. The Tower of London has many fine examples -- as you're leaving Tower Hill tube station, glance up and to the left just before the exit. Pinned to the wall is a colourful if dusty sign advertising a long-forgotten wireless service named Phonepoint. Alongside its competitors Zonephone, Callpoint and Rabbit, this was part of a very high profile phone system called CT-2: flexible and affordable, it was going to bring digital mobile phones to the masses years before GSM. The masses thought otherwise. Launched in 1989 by high-powered consortia of banks, national telcos and equipment manufacturers, the system was dead by 1994.
CT-2's parallels with Wi-Fi are striking. The ideal user would have a base station at home, tied to their ordinary phone line. They'd use telepoints -- the CT-2 term for hot spots -- on their travels, and another in the office would link them to the work system. Telepoints appeared at stations, busy shopping centres and airports. There were a number of rather natty phones to choose between and, early teething problems notwithstanding, the system worked as planned.
It failed for a number of reasons. The most obvious was that you couldn't get incoming calls when you were on the move: there was no way to route a call to a telepoint. People carried pagers, and when they were bleeped they'd try and find somewhere to call back from. Telepoints weren't that common, though: the punters soon realised it was far cheaper and more reliable to ditch the phone, keep the pager, make like Clark Kent and leap into the nearest phone box.
Then there was compatibility between networks -- rather, there wasn't. Each network stood proud and alone, so if you had a Rabbit phone and were passing a Callpoint telepoint you might as well have been carrying a brick. CT-2 had been designed as a proper open standard, and there were no technical reasons why companies couldn't share infrastructure, but telco chief executives in the 1980s would as soon share a toothbrush. As for the work/home mix: people had phones in both places already. Why bother? As soon as cellular phones appeared in any numbers, the disadvantages of CT-2 became fatal.
The lessons are: people won't carry anything that merely duplicates the functions of something they have already; they have no tolerance for incompatibilities that are the result of corporate rivalry; and they all own the Dead Kennedys record Give Me Convenience Or Give Me Death.
Let's look at the proponents of Voice over IP combined with wireless networking. Fantastic opportunity for convergence, they cackle, especially if they're Cisco with fingers firmly in both pies. Hot spots are sweeping the world. Low call costs. Network access. From a user's viewpoint, the picture is less rosy.
For a start, there's the small matter of what to wear. We all have mobiles already. Will phone companies put Wi-Fi into their products? There's a lot of resistance from the carriers, and currently little advantage to the users. You can carry a PDA with Wi-Fi and a voice stack, sure, but it won't work anything like a mobile phone: you can sit in a hotel lobby with a headset plugged into your laptop, but you'll look several thousand times less cool than someone with the latest, shiniest mobile glittering away in their hands.
Companies don't care about cool: they care about cash and that's about it. You can buy a Wi-Fi-only phone and configure your work network to support it -- this looks good on paper and saves organisations with big campuses big money. The users, who have to carry around a work phone as well as their personal mobile, may not be so enthused. If mobile phones, like PCs, had a standard expansion bus that meant third parties could add networking independently of the original manufacturers, then the dynamics would be entirely different. There is no such standard, nor is there any sign that the industry wants one: we know who inherited the CT-2 toothbrush.
Can your friends find you when you're logged on? Instant messaging works, but that is a morass of incompatible standards: I already run three IM clients on my PC, and I don't want to do that on the mobile as well. Can a mobile client switch instantly, silently and efficiently between hot spots run by different companies? Ask anyone who's spent a day swearing at network card A's unwillingness to talk to access point B, downloading drivers and Googling for the true meaning of inscrutable error messages. As for billing: right. All these things must work perfectly before public voice over Wi-Fi has a hope. That's not enough.
For voice over Wi-Fi to be a serious threat to the cellular system, it can't be a me-too system. The things that make it attractive to corporates won't cut it with private users, and almost anything you can do with Wi-Fi and voice you can do with cellular networks -- where the incumbents are well-funded, dug in like rattlesnakes and more than willing to fight.
Voice over Wi-Fi needs that old-fashioned idea, the killer application. And that has to use the one thing that Wi-Fi has that 2G or 3G phones do not -- megabit broadband. As any ISP will tell you, that's shorthand for digital media delivery. When my Wi-Fi phone in Wichita will play my hi-fi sound files from my home server in London, then I'll listen. Until then, I'll carry on taking in the sights of Old London Town, the graveyard of a million good ideas done badly.