If you're ever asked to forecast the way computing will develop, offer to look three to five years ahead.

It's a good, safe time frame because if you're right then people may just remember your prediction when you remind them how clever you are, and if you're wrong it's very unlikely anyone will think to point it out.

Trying to anticipate significant developments for the coming year is a lot harder, perhaps because the tendency is to overestimate the impact of the few obvious trends and miss the slow-burn developments that are on the verge of going mainstream and changing the way we see the world.

For example, last year I wrote "we are building our lives around the network and the things it makes possible, and 2006 marks the year in which this became a sensible and indeed rather normal thing to do rather than something that marked you out as a geek".

While it's true that Facebook and other social network sites went mainstream, they are still not as widespread as the sometimes breathless coverage would make you think.

Wireless world

It's the same with my prediction about phones, when I argued that "we're going to see smartphones and mobile access finally come into their own, as the devices live up to the earlier promise and the networks finally realise that treating handsets as network nodes makes a lot more sense than acting like they are mobile phones with added data services".

This has started to happen, but it's going to take a long time before we're all surfing the wireless web from smart mobiles.

The iPhone has accelerated the process begun by Symbian, and the rollout of Google's Android and open source phones like OpenMoko may help, but it will be a few years before the devices are completely freed from reliance on the network.

One facet of mobile internet access may change quite fast, however.

Buying wifi by the hour in cafes or on trains is expensive and tedious, and the widespread availability of 3G data cards for laptops on fixed monthly rates could hasten the demise of the pay-per-use services.

It may force the operators to do more deals to offer free access like the one between The Cloud and McDonalds, and I wouldn't be surprised to see free wireless in Starbucks by the middle of the year.

At home we'll see faster broadband services being delivered over cable, and ADSL providers will try to keep up.

BT and the other telcos will complain loudly about not being able to afford the investment needed to upgrade the local loop between exchanges and homes, but if the government keeps its nerve and refuses them tax breaks I think we'll find that the money is there after all.

The potential revenue from fast broadband networks are just too great to pass on.

Network power

We'll also see better screens.

The multi-touch interface that Apple has built into the iPhone and the iPod Touch will be used on bigger devices, perhaps giving the tablet PC a new lease of life after years in which it has struggled to escape the taint of Microsoft's over-enthusiastic marketing.

And there will be more and better location-based services. The newest version of Google Maps for mobile does a pretty good job of figuring out where you are without GPS by using the cell network, and we may see this information used to tag photos and blog posts.

No doubt Facebook will seize the opportunity to offer an "I'm here" map-based addition to your profile, whether you want it or not.

Together these changes amount to more of the same, offering us easier, simpler, cheaper and faster access to the network.

But next year's real shift will be more subtle and have much greater long term impact.

At the moment most of the computing we do is local, and programs run on our laptops or desktops.

This is starting to change, and in 2008 we will see more and more processing moving away from the user and into large data centres which serve many different organisations.

Gradual shift

The change is described in Nick Carr's new book The Big Switch, where he argues that computing power is becoming a utility.

He believes that instead of owning our own processors we will soon be renting time on large systems run by the likes of Google and Amazon, not just for storing data but also for running code.

Carr sees strong parallels between the way electricity generation shifted from local generators in factories to a national grid providing voltage differences where they are needed and the move from local to central processing.

In the world Carr describes most processing takes place in "the cloud", and the computers we actually use will manage the interface and the communications, but do little of the real work.

It is a compelling vision, though not without its problems.

Computing is not a simple service like electricity, and it's not clear that we can solve the administrative problems needed to have business-critical services hosted remotely.

Moving everything onto the network may appeal in the rich countries of the industrialised world but offers little to rural India or sub-Saharan African countries.

And there are massive security and data management issues to be solved.

Even so, the potential benefits are too great to be ignored, and we're likely to see a range of services go live next year that will, if successful, take us closer to the cloud computing model.

It will not be an overnight shift, but when we look back in a decade or so I think we'll see 2008 as the year things started to change.

Bill Thompson is an independent journalist and regular commentator on the BBC World Service programme Digital Planet.