I've described elsewhere the idea of "swarms" - spontaneously federating devices and software services connecting over networks. Some people are now describing this concept as "wireless Web services," extending the group of ideas now being called services-on-demand.

As usual, the computer industry is keen to address the details of protocols and connections, but is leaving until later the real, people-centric issues that technologies can't solve.

Possibilities
Imagine the possibilities once we can have the services not only offered to us on demand via the nearest device but also have them include both our current context and our online profile. I'll be able to tell my car who to start for and how each person may use it ("yes, you can borrow it, but don't drive home if you're drunk, and stay under 50"). As my current meeting ends late, my PDA will offer suggestions for how to reschedule the rest of the day's meetings and travel. The just-in-time production line will be able to respond to all the inputs, including current demand in the retail stores. All sorts of ideas that we thought were the domain of intelligent agents will start to become a reality.

But hold on a moment. Can it all really happen like that? For some time now my thesis has been that we already have, or can imagine, all the technology we need to build these spontaneously federating solutions. The real issues lie elsewhere.

Standards
The first priority has to be open, loosely coupled standards, for content as well as infrastructure. In the area of standards, it's clear that we've made progress over the last decade. The Web browser gives me a standard space to interact with remote computers. GSM gives me a mobile phone that works worldwide. But we're not there yet with Web services, let alone with wireless Web services - none of the basic technologies is actually an open, royalty-free standard today. Discussion over the actual XML content of the transactions, although in progress at ebXML, is in its youth and under-supported by the key vendors. And peer-to-peer ideas tend to be neglected altogether.

If we've learned one thing from a decade of the Web, it's that the massively connected mesh demands open, loosely coupled standards - "open" in the sense of available to anyone to develop with or use without fee; "loosely coupled" in the sense of tolerating extension without opening a path to proprietary lock-in; "standards" in the sense that a democratic process offers every affected developer and user the chance to be involved in each change.

For services-on-demand (both Web services and P2P services) - to be offered to the mobile user, we'll need a serious commitment to openness and standards that extends beyond the mode of merely reconciling infrastructure for proprietary content that we're seeing at the moment from some.

Business Models
Secondly, we'll need some careful progress made on business models. Web services void the only business model that "free" services on the Web ever had - that of exploiting eyeballs. Without advertisements, many information providers will have to look elsewhere for a revenue model. So far as European telcos are concerned, the urgent need to validate their investments in 3G licenses may lead to progress here. We may see commercial-quality service federations created that are free at your desk but part of the service plan on the move - as long as we can overcome the contractual, legal, and intellectual property problems.

Logistics
The procedural barriers may, thirdly, prove to be the killers. Realistically, no one vendor or service provider will ever be able to provide everything you need - it will take a community. Building communities is the lifeblood of progress on the Web and has underpinned Web standards (W3C is essentially an expert community), open source development (from pioneers like Linux and Apache to commercial foundations like NetBeans and OpenOffice), and technology evolution. But the application space gets more complex once we go mobile. What about antitrust laws? What about intellectual property ownership? What about contractual protection?

And then there are the international issues - local languages, differences in jurisdiction, and so on. Can my swarm continue to work across state lines? How about across international boundaries? Or with local services in a country with a different local language?

Opportunity
After reading this far, you might think I'm pessimistic about services-on-demand being offered to the mobile user. But, on the contrary, I'm very optimistic. In the mobile space, in the absence of the monopolistic pressures that have impacted the PC market, we've seen extensive industry agreement. Up to this point, we've seen agreement on GSM and Java technology deliver a level playing field to the network operator and to the developer.

The plumbing is in place. But people and business issues have always been the linchpins in making new technologies real in the mass market. It's true for wireless Web services-on-demand as well.

• Figure 1